Forest
tree
improvement
Report on the
FAO/DANIDA Training Course
on Forest Tree Improvement
Merida, Venezuela
January - February 1980
The designations employed and the presentation
of material in this publication do not imply the
expression of any opinion whatsoever on the
part of the Food and Agriculture Organization
of the United Nations concerning the legal
status of any country, territory, city or area or
of its authorities, or concerning the delimitation
of its frontiers or boundaries.
M-31
ISBN 92-5-100943-0
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted in any form or by any means,
electronic, mechanical, photocopying or otherwise, without the prior
permission of the copyright owner. Applications for such permission,
with a statement of the purpose and extent of the reproduction, should
be addressed to the Director, Publications Division, Food and Agriculture
Organization of the United Nations, Via delle Terme di Caracal I a, 00100
Rome, Italy.
> FAO 1M6
- iii -
ABSTRACT
The FAO/DANIDA Training Course on Forest Tree Improvement was held in Venezuela
from 14 January to 2 February 1980. The Course was organized by the FAO Department of
Forestry in collaboration with the Government of Venezuela, the University of the Andes,
and the Institute Latinoamericano de Investigacion y Capacitacion. The Course was financed
with funds made available by the Danish International Development Agency (DANIDA). It
was attended by 19 experts from 17 countries of Latin America: Argentina (1), Bolivia (1),
Brazil (1), Chile (1), Colombia (1), Costa Rica (1), Cuba (1), Dominican Republic (1),
Ecuador (1), Guatemala (1), Honduras (1), Nicaragua (1), Panama (1), Paraguay (1),
Peru (1), Uruguay (1) and Venezuela (3).
The Course consisted of two weeks of lectures and practical demonstrations in
Merida and one week study trip to the states of Barinas and Monagas in western and
eastern Venezuela respectively.
Lectures included the following topics: tree improvement in relation to national
forest policy; elements and principles of genetics; conservation and rational use of
forest genetic resources; collection and handling of forest seeds; storage, testing and
certification of forest seeds; experimental designs; statistical interpretation of test
results; species and provenance trials; selection and management of seed stands;
selection of forest trees; vegetative propagation methods; controlled crossing systems
and designs; establishment and management of seed orchards; progeny trials; genotype/
environment interaction; breeding for disease resistance; strategies for tree development
programmes; economic considerations of forest tree breeding programmes.
- V -
TABLE OF CONTENTS
Page
1 . Introduction 1
2. Organization and Conduct of the Course 1
3. Conclusions 1
4. Acknowledgements 2
ANNEXES
1. List of lecturers, support personnel and participants 4
II . Course programme : theory 8
III. Course programme: study trip 9
IV. Lectures : 10
Tree improvement in Relation to National Forest Policy
and Forest Management (R.L. Will an) 11
Elements and Principles of Genetics (W.H.G. Barrett) 15
Principles and Strategies for the Improved Use of
Forest Genetic Resources (C. Palmberg) 23
Sampling in Seed Collection (C. Palmberg) 41
Collection and Handling of Forest Seeds (C. Palmberg
and G.H. Melchior) 46
Storage, Testing and Certification of Forest Seeds
(B. Ditlevsen) 60
Experimental Designs (B. Ditlevsen) 75
Statistical Interpretation of Test Results (B. Ditlevsen) 86
Species and Provenance Trials (R.L. Willan) 103
Seed Stands (M. Quijada) 112
Selection and Management of Seed Stands with Special
Reference to Conifers ( W.H.G. Barrett) 116
Selection and Management of Seed Stands: Hardwoods
(C. Palmberg) 122
Selection of Forest Trees (M. Quijada R. ) 124
Quantitative Genetics: General Principles and Practical
Applications in Forest Tree Improvement (B. Ditlevsen) 131
Vegetative Propagation Methods (M. Quijada R. ) 140
- vi -
Page
Controlled Crossing Systems and Designs (B. Ditlevsen) 148
Seed Orchards (W.H.G. Barrett) 160
Progeny Trials (M. Quijada) 168
Genotype Environment Interaction (M. Quijada) 174
Breeding for Disease Resistance (C . Palmberg) 178
Economic Considerations in Forest Tree Breeding Programmes
(B. Ditlevsen) . . . ."7 187
Planning and Strategies of a Tree Development Programme
(C. Palmberg, D.K. Paul & R.L. Willan) L 197
Some Aspects of the Problem of Genetic Improvement of
Hardwood Species Native to Venezuela (M. Quijada) 213
V. Study trip notes
1. Flowering, Seed Production and Artificial Pollination
in Bombacopsis quinata in Venezuela 216
2. Plantation Programme in Llanos Orientales, Venezuela 219
3. Activities and Human Resources needed for the CVG
Plantation Programme, Uverito 220
4. Calendar of CVG Activities , Uverito 222
VI. Country Reports 223
VII. Bibliography 266
1. Publications distributed to participants 266
2 . Other publications of interest 267
VIII. Certificate of participation for participants 271
Opening of the Course. C. Palmberg (FAO) ; Dr. P. Rinc6n Gutierrez (Rector, University of
the Andes); Dr. R. Chalbaud Zerpa (Governor of the State of MSrida) ; and J.R, Corredor
Trejo (Dean, Dept, of Forestry Science, ULA) .
Opening of the course. Dr. R. Chalbaud Zerpa (Governor of the State of MSrida) ;
C. Palmberg (FAO Co-Director of the course); B. Ditlevsen (DANIDA Co-Director of the
course); A. Luna Lugo (ULA); and W.H.G. Barrett (International Director of the course).
1. INTRODUCTION
The FAO/DANIDA Course on Forest Tree Improvement was held in Venezuela from
14 January to 2 February 1980, at the kind invitation of the Government of Venezuela.
The Course was organized by the Department of Forestry of FAO with funds made available
by the Danish International Development Agency (DANIDA). With the technical collabora-
tion of the University of the Andes in MSrida and the support of the Institute Latino-
american de Investigacion y Capacitacion, IFLAIC.
The training course was one in a series as part of the FAO programme for forest
tree improvement. Earlier courses in the series were held in Denmark in 1966, in Kenya
in 1973 and in Thailand in 1976.
The objective of the course held in the Latin American region, and which was
intended for professional foresters, was to provide participants with a theoretical and
practical background in present-day forest tree improvement, and to promote contact
between foresters and forestry institutions in the region.
Nineteen participants from 17 Latin American countries participated in the training
course (see Appendix I).
Dr. W.H.G. Barrett of Argentina was the International Director of the course.
Dr. M. Quijada R. of the Institute of Silviculture, University of the Andes, acted as
national Co-Director. Dr. B. Ditlevsen of Denmark acted as the DANIDA Co-Director and
C. Falmberg was the FAO Co-Director.
2. ORGANIZATION AND CONDUCT OF THE TRAINING COURSE
The first two weeks of the course covered both theoretical and practical aspects
of forest tree improvement. The third week was devoted to a study trip to the states of
Barinas and Monagas in western and eastern Venezuela respectively.
Appendixes II and III list the detailed programme of the course.
This training course benefitted from the experience of earlier courses.
With respect to earlier courses, this training course placed heavier emphasis on
experimental designs and statistical evaluation of trials. Tropical hardwood tree improve-
ment techniques were also included where possible.
The close relationship between national forestry policy and tree improvement ^
grammes was stressed throughout the training course. An attempt was made to give partici
pants a broad perspective of the tree improvement issue. The various prerequisites of
justifying a new tree improvement programme such as available land and a solid reforesta-
tion progranme were also stressed.
Before the course the instructors prepared written lecture notes which were dis-
tributed to participants on their arrival in Mfirida. Participants also received certain
publications and papers of fundamental importance of general interest. The lectures are
found in Appendix IV. A bibliography, including the publications distributed to partici^
pants, is found in Appendix VII.
Bef9re their arrival in MSrida, participants filled out a questionnaire on the
current plantation and forest tree improvement situation in their respective countries.
This information formed the basis of the brief talk by each participant based on the
"Country Reports" in Appendix VI.
3. CONCLUSIONS
The course offered expertise which will be applied and extended by participants
upon their return to their respective countries. This aspect of the course should cer-
tainly not be under-rated, but the opportunity for 19 participants representing 17
- 2 -
countries to meet and exchange information and experiences and discuss common problems
with experts from other countries was also, and unquestionably, a unique and valuable
opportunity.
This was a particularly opportune moment for a course of this type because most
of the Latin American countries are intensifying plantation activities with an eye to
meeting their wood and fuelwood requirements, and are including the introduction of
exotic species and selection and improvement programmes in their activities.
The study trip at the end of the course was highly important and very valuable.
Participants were given the opportunity to see problems which can occur in real life
situations and to discuss possible compromises and ways of solving these problems.
It should be stressed that the success of the course was largely due to the
interest shown by participants. This made it possible to overcome certain difficulties
which arose at the outset, such as the range of technical expertise met at the meeting.
The camaraderie and good humour which were a permanent feature of the workshop helped
create a propitious, inspiring atmosphere for the activities of the course.
4. ACKNOWLEDGEMENTS
FAO wishes to express its appreciation to the Government of Denmark, who sponsored
the course through the Danish International Development Agency, and to the Government of
Venezuela who offered to host the course. In addition, thanks are due to the Department
of Forestry Sciences of the University of the Andes, Merida, for its very valuable admini-
strative and professional support and to the Consejo de Desarrollo Cientlfico Humanistic
(CDCH) for its financial support; the Institute Latinoamericano de Investigacion y
Capacitacion, IFLAIC, which made its lecture rooms and libraries available to the course;
the Corporacion Venezolana de Guayana, CVG, and the Compaftia Nacional de Reforestacion,
CONAKE, through whose cooperation the study trip was made possible; the Laboratorio
Nacional de Productos Forestales, LABONAC, and the Training Centre of the University of
the Andes for their collaboration throughout the course.
Special thanks are due to the Governor of the State of Mgrida, Dr. Reinaldo
Chalbaud Zerpa, who contributed substantially to the success of the course through his
personal interest, ensuring that participants were thoroughly briefed on the traditions
and customs of Venezuela and of Mirida.
Finally FAO wishes to thank all lecturing and support staff and expresses its
appreciation for the interest of participants and also extends thanks to all those who,
through their interests and efforts, helped to make this a successful training course.
- 3 -
COURSE PARTICIPANTS AND INSTRUCTORS;
1st Row, left to right: M.E. Qulnteros (Paraguay); J.A. Enricci (Argentina);
A. Gonzalez (Cuba); A. Gomez de Fonseca (Brazil); A. Martinez (Panama); G. Moreno (Chile).
2nd Row, left to right: J. Morales (Venezuela); A, Zapata (Venezuela); G.H. Raets (IFLAIC) ;
J. Campos (Venezuela); R.A. Rodriguez (Dominican Republic); M. Quijada (Venezuela);
R. Miliani (Venezuela); C. Palmberg (FAO); W.H.G. Barrett (Argentina); D.J. Moreno
(Bolivia); T. Quintini (Venezuela); R. Escudero (Uruguay); B. Ditlevsen (Denmark);
P.E. Silvade la Maza (Nicaragua); O.V. Anleu (Guatemala); H.E. Carrillo (Peru);
D. Villalobos (Honduras); R. Valcarcel (Brazil); G.E. Porras (Costa Rica); A. Ramirez
(Venezuela); C.C. Castillo (Venezuela); 0, Carrero (Venezuela); A. Copete (Colombia);
C. Linares (Venezuela).
Institute of Silviculture , University of the Andes, Mgrida.
- 4 -
Appendix I
FAO/DANIDA
TRAINING COURSE
FOREST TREE IMPROVEMENT
LECTURERS AND SUPPORT PERSONNEL
Dr. Wilfredo E.G. Barrett
Fiplasto Forestal
Maipu 942 - Piso 21
1340 - Buenos Aires
Argentina
Dr. Bjerne Ditlevsen
Skovstyrelsen
Strandvejen 863
DK-2930 Klampenborg
Denmark
Dr. Marcelino Quijada R.
Seccion de GenStica Forestal
Institute de Silvicultura
Universidad de Los Andes
MSrida
Venezuela
Miss Christel Palmberg, M.F.
Forest Resources Division,
Department of Forestry
FAO, Via delle Terme di Caracal la, 1
00100 Rome,
Italy
Ing. Herman Finol U.
Universidad de Los Andes
Facultad de Ciencias
Fores tales, Via Chorros de
Milla, Merida
Venezuela
Ing. S. Gimenez Fonseca
Universidad de Los Andes
Facultad de Ciencias
Fores tale s, Via Chorros de
Milla, Merida
Venezuela
Dr. Oton Holmquist
Universidad de Los Andes
Facultad de Ciencias Fores tales,
Via Chorros de Milla,
Merida
Venezuela
Ing. L. Rodriguez Poveda
Universidad de Los Andes
Facultad de Ciencias Fores tales,
Via Chorros de Milla,
MSrida
Venezuela
Dr. G.H. Raets
Institute Latinoamericano
de Investigacidn y
Capacitacion, Apart ado 36
Merida
Venezuela
- 5 -
PARTICIPANTS
ARGENTINA
Ing Juan Andres Enricci
Estacion Forestal Trevelin
Casilla Correo No. 17
Revel in - Chubut (9203)
BOLIVIA
Ing Deimer Jesus Moreno
PRONAPLAF
Casilla Correo No. 209
Avaroa No. 637
Sucre
BRASIL
Ing Agostinho Gomes da Fonseca
I.B.D.F./D.Pq. - Ed.
Palacio do Desenvolvimento
13 Andar - SBN.
Brasilia - D.F.
COLOMBIA
Ing Alejandro Copete Perdomo
INDERENA
Institute Nacional de los
Recursos Naturales Renovables
Calle 26 No. 13 B-47
Bogota
CUBA
Lie. Anibal Gonzalez Roque
Seccion de Genetica
Centro de Investigaciones Forestales
Calle 174 No. 1723 E/ 17B y 17C
Siboney - Marianao
ECUADOR
Ing Jaime Narvaez
Arosemena Tola No. 452
Urb. Borja Yerovi Sector 32
Quito
CHILE
Ing Gustavo Moreno Diaz
Corporacion Nacional Forestal
Centro de Semi lias
Casilla 5
Chilian
COSTA RICA
Ing Guillermo Enrique Porras Sandoval
Direccion General Forestal
Ministerio de Agricultura y Ganaderia
San Jose
DOMINICAN REPUBLIC
Sr. Ramon Agustin Rodriguez R.
Direccion General Forestal
Centro de los Heroes
Apartado Postal 1336
Santo Domingo
GUATEMALA
Sr. Osman Vinicio Anleu
BANSEFOR
Institute Nacional Forestal
7 a Av. 7-00 Zona 13
Guatemala Ciudad
HONDURAS
Sr. Angel Danilo Villalobos Nuftez
Seccion Mejoramiento Genetico
Escuela Nacional de Ciencias
Forestales, Apdo No. 2
Siguatepeque
PANAMA
Lie. Aristides Martinez Montilla
Seccion de Mejoramento Genfitico
Direccion Nacional de Recursos
Naturales Renovables
Paraiso, Panama S
NICARAGUA
Lie. Pedro Eloy Silva de La Maza
Seccion de Invest igacion Forestal
Institute Nicaraguense de Recursos
Naturales y del Ambiente, I.R.E.N.A.
Managua
PARAGUAY
Ing Martin Eugenio Quinteros Doldan
Servicio Forestal Nacional
Centro Forestal Alto Parana
Cd. Strossner
- 6 -
PERU
Ing Hugo Edgar Carrillo Vargas
Direction General Forestal y de Fauna
Regi6n Agraria IV
Prolongaci6n Raimondi S/N
Huaraz
VENEZUELA
Ing Oswaldo Carrero
Centre de Invest igaci5n Forestal
Companla Nacional de Reforestacion
CONARE
Apdo. No. 264
Maturin/Edo. Monagas
Ing Roberto Miliani
Estaclon Experimental de Semi-lias
Fores tales, M.A.R.N.R.
El Limon
Maracay/Edo. Aragua
URUGUAY
Ing Rafael Escudero Rodriguez
Facultad de Agronomla
Departamento Forestal
Universidad de la Republica
Avenida Garz6n 780
Montevideo
Ing Tomas Quintini
Sub-Gerencia Forestal
Corporaci6n Venezolana de
Guayana, CVG
Centro Comercial Los Olivos
Puerto Ordaz/ Edo. Bolfvar
OBSERVERS
Lie. Carmen Cecilia Castillo
Division de Desarrollo Agrfcola, CVG
Edificio La Estancia, Piso 13
Chuao, Caracas
Sr. Jos Campos
Sub-Gerencia Forestal, CVG
Centro Comercial Los Olivos
Puerto Ordaz /Edo. Bolivar
Ing Jairo Morales
CONARE, Programa Coloradito
Apartado 196
El Tigre/Edo. Anzoategui
- 7 -
Bombacopsis quinata, genetic variation in the occurrence of buttresses.
- 8 -
- 9 -
Appendix III
27 January
28 January
29 January
30 January
31 January
1 February
2 February
COURSE PROGRAMME
(Study trips)
Trip by bus to Barrancas. "El Irel 11 experimental station.
Controlled pollination demonstration on Bombacopsis Quinata.
Caemital experimental forest: trials at "El Irel station".
Trip to Puerto Ordaz, "El Merei" forestry site.
Visit to CVG plantations, nursery and trials in Uverito.
Visit to CONARE plantations, nursery and trials in Chaguaramas.
Trip to Puerto Ordaz, visit to "La Llovizna" park and SIDOR plant
(Siderurgica del Orinoco).
End of course.
"Perils" during the study trip
- 10 -
Appendix IV
LECTURES
(Contents)
Page
Improvement of forest trees in relation to national forestry policy and
management of forest trees (R.L. Willan) 11
Elements and principles of genetics (W.H.G. Barrett) 15
Principles and strategies for the improved use of forest genetic
resources (C. Palmberg) 23
Sampling in seed collection (C. Palmberg) 41
Collection and handling of forest seeds (C. Palmberg & C.H. Melchior) 46
Storage, testing and certification of forest seeds (B. Ditlevsen) 60
Experimental designs (B. Ditlevsen) 75
Statistical Interpretation of trial results (B. Ditlevsen) 86
Species and provenance trials (R.L, Willan) 103
Seed stands (M. Quijada) 112
Selection and management of seed stands with special
reference to conifers (W.H.G. Barrett) 116
Selection and management of seed stands: hardwoods (C. Palmberg) 122
Selection of forest trees (M. Quijada) 124
Quantitative genetics: general principles and practical
application in forest tree Improvement (B* Ditlevsen) 131
Methods of vegetative propagation (M. Quijada) 140
Controlled crossing systems and designs (B. Ditlevsen) 148
Seed orchards (W.H.G. Barrett) 160
Progeny trials (M. Quijada) 168
Genotype /environment Interaction (M. Quijada) 174
Breeding for disease resistance (C. Palmberg) 178
Economic considerations on forest tree Improvement programmes (B. Ditlevsen) ... 187
Planning and strategies of a tree improvement programme
(C. Palmberg, D.K. Paul & R.L. Willan) 197
Some aspects of the problem of genetic Improvement of
hardwood species native to Venezuela (M. Quijada) 213
- 11 -
Tree Improvement in Relation to
National Forest Policy and Forest Management
R.L. Willan
FAO
CONTENTS
Page
Introduction 11
Prerequisites for a tree improvement programme 12
Defining objectives 12
Constraints 13
Tree Improvement in relation to management 13
Summary 14
Bibliography 14
INTRODUCTION
The justification of any national programme of tree improvement and the amount of
resources devoted to it should be largely determined by the forest policy of the country
(Keiding 1974). Forest policy, in turn, must be related to national development plans.
Basic questions which need to be asked are (1) How significant is forest in the
nation's economy? and (2) How significant is plantation forestry in the forestry sector
as a whole? At one extreme is the case of a small, highly populated country in which
there is no room for extensive forestry, and therefore no place for tree improvement.
Singapore and Malta are examples. At the other extreme are countries with low population
density and large areas of natural forest capable of being naturally regenerated and of
providing the country's needs in the foreseeable future. Gabon and, until recently,
Kalimantan and Indonesia may be examples. In such countries there is a case for a simple
silvicultural technique designed to retain a proportion of the best trees as seed-bearers
and, without doubt, for the establishment and protection of strict natural reserves
within the forest, but not for a long-term programme of tree breeding. In regions where
the crucial protective function of the forest in steep topography precludes commercial
harvesting and hence reforestation, there is likewise no scope for tree breeding.
Few countries can now afford to rely entirely on natural forests. In most the
pressure on land is becoming steadily greater and, if the requisite raw materials can be
produced from a much smaller area or in a shorter period by using intensive high-yielding
plantation methods, foresters have a responsibility to employ these methods. Concentrated
plantations lead to indirect economic and social benefits through the development of forest
industries and opportunities for employment. Accordingly, plantations are being established
on an increasing scale throughout the world. In practice the roles of natural and man-made
forests should be considered as complementary, with natural forests fulfilling essential
protective and cultural functions and providing certain special categories of wood, such
as high quality material for cabinet making or veneering, while plantations supply an in-
creasing proportion of requirements for constructional timber, utility sawn wood, wood
based panels and pulp (Hughes and Willan 1976).
Generally speaking, the prerequisite for a tree improvement programme is plantation
forestry. As soon as seed is collected, plants raised and artificially cultivated, there
is a chance to select and improve. Thus, it will be convenient to start the consideration
of tree improvement in relation to afforestation and reforestation programmes (Keiding
1974). This is why country statements which participants have been requested to complete
set the information on national tree improvement programmes in the context of the national
environment, the national forest policy and national programmes on afforestation and re-
forestation.
- 12 -
PREREQUISITES FOR A TREE IMPROVEMENT PROGRAMME
Prima facie Even though there may be a case for plantation forestry and tree breed-
ing in a country, expenditure of funds and effort demands certain prerequisites :-
A. The planting programme
(1) Availability and control of land. The large investment involved in planta-
tion forestry can be justified only if there is an assurance that forestry will remain the
object of land management for at least one rotation. Even with "fast-growing species"
this is likely to extend for one or more decades. And the managerial authority must have
full control of the land throughout the period. Excellent security of tenure may exist
if the plantations are on government-owned forest land included in a national land usage
plan by which continuity of management is promised for some years ahead. Land which is
fragmented among numerous small private owners is usually unsuitable for plantation fores-
try. On the other hand, tree breeding may have a part to play in the provision of trees
for diffuse planting in agrisilviculture, provided that the farmer is convinced of the
value of the product and of the need to protect and manage the trees.
(2) Scale of operations. No matter how great the gains to be derived from tree-
breeding, the basic minimum costs of a small research unit need to be spread over an
adequate area if they are to pay for themselves. For example a unit costing $100 000 a
year and producing improved seed capable of yielding an increase in discounted product
value of $100 per ha per year, would more than pay for itself on a 10 000 ha a year
programme but could not be justified for a 100 ha a year programme.
(3) Availability of markets. There needs to be reasonable assurance of markets
for plantation produce, either within the country or through export. Not only must the
markets exist, but they must be within economic distance. Plantations, even on* high-
yielding sites, may be uneconomic if transport costs are crippling.
B. The tree improvement programme
(4) There must be reasonable assurance, e.g. in a written policy statement by the
financing authority, that staff and funds will be provided to the tree improvement pro-
gramme on a continuing basis* As Zobel (1969) has stated: "Will I have the backing in
funds, facilities and manpower to do a decent job? If not, then don't start! A half-
hearted programme, poorly done, will only sour people on forestry and its potentials."
If a trained tree breeder is not already available from within the country, provision to
train one must be made from the start.
(5) Assessment of technical information available from elsewhere. Results of
research in other countries with similar environments may reduce, if not eliminate, the
need to start a national tree improvement programme de novo. For small countries with
modest planting programmes, a regional research unit may provide the same results as
several national programmes and at a lower cost. Examples of regional tree improvement
programmes are those which operated in East Africa during the 1960s and ' 70s and CATIE
now operating for Central America. Even where a country's planting programme is large
and is carried out on a unique range of sites, International exchange of information and
of genetic material can do much to avoid duplication and concentrate research on solving
the most important problems or exploiting the most promising opportunities.
DEFINING OBJECTIVES
Having established that the basis for a national tree improvement programme exists,
the policy maker needs to define its objective as clearly as possible. There are often
good reasons for specifying more than one objective. In view of the long-term nature of
forest crops and, in comparison, the speed with which technological preferences and markets
can change, there is often a strong case for maintaining the maximum flexibility in the
objectives. Flexibility is especially important with rotations longer than 15 years
(Hughes and Willan 1976). However, there are limits to the flexibility of programmes
and species which should be recognized from the start. A high density eucalypt is likely
to be superior to a low density eucalypt for fuelwood but inferior for short-fibre pulp-
wood. In the southern USA separate seed orchards are planted for high-density and
- 13 -
low-density wood in pines. Thus different purposes may sometimes be served best by grow-
ing separate crops using separate selection criteria rather than by attempting to grow
a "multi-purpose 11 crop which serves neither purpose adequately. If more than one objec-
tive is specified and in case resources prove inadequate to achieve all objectives - which
is almost inevitable - it is important to assign relative priorities.
The objectives defined must be based upon immediate, short term and long term
requirements of the relevant afforestation programme. Constraints on the achievement of
objectives also must be taken into account at the time of definition, and should indicate
the limits of the resources (staff, facilities and finance) within which the project will
be required to operate (Hughes and Willan 1976).
CONSTRAINTS
Tree improvement objectives can only be achieved within existing biological human
and financial constraints. In some cases these are so limiting that the objectives are
impossible of attainment and it becomes the duty of the tree breeder to convince the
policy maker to set more realistic objectives. Examples in which limiting conditions
could force a change in the objectives might be:-
(1) Attempts to grow high-yielding plantations in semi-arid conditions. Species
of provenances may be fast-growing or drought-resistant, but rarely both. Any improve-
ment from tree breeding is likely to be small and slow. More impressive improvement may
be obtained by shifting the afforestation scheme to a more humid zone or by introducing
irrigation.
(2) Attempts to start a tree improvement programme in an introduced species,
using small plantations of narrow or unknown genetic base. In such cases further intro-
ductions and provenance trials are more urgent than individual selection in the existing
plantations.
(3) Attempts to use an exotic species of which foreign seed supplies are inade-
quate and which fails to produce a seed crop in the new environment.
(4) Attempts to improve traits which are known to have very low heritabilities.
(5) The time-scale imposed by the period taken from germination to seed production
in forest trees is often a serious constraint. If somewhat improved seed is needed in
five years and the species takes 10 years to produce seed in a seed orchard, then estab-
lishment of a seed orchard now cannot achieve the objective. Alternatives, such as heavy
improvement fellings in commercial stands, or a change of strategy, such as development
of vegetative propagation methods, must be considered.
TREE IMPROVEMENT IN RELATION TO MANAGEMENT
Tree improvement is simply one tool available to forest management and cannot be
considered in isolation. Research on tree improvement should therefore be closely inte-*
grated with other research, e.g. on soils and site assessment, on establishment techniques,
on spacing and thinning and on wood quality. National forest policy must therefore in-
clude provision for the integration of all sectors of forest research and for the rapid
multiplication and introduction into forestry practice of improved material.
Tree improvement results may modify forest management. For example better trees,
more uniform and more disease resistant, make possible a wider initial spacing which
reduces costs while still achieving the desired final stocking (Zobel 1969). Breeders
can produce genotypes which respond well to intensive site preparation and fertilization
and other genotypes which will tolerate poor soils and no fertilizers. Within a single
country there are likely to be a number of different site types to be planted and the tree
breeder therefore needs to develop a range of genotypes adapted to the various sites.
Technological developments may alter breeding priorities, e.g. the invention of a new
pruning device leading to a major reduction in pruning costs may reduce the need to breed
for small branches and early self-pruning, or a revolutionary new pulping process might
reduce the Importance of wood quality or wood uniformity.
- 14 -
SUMMARY
(1) Tree improvement programmes must be closely related to the objectives and
priorities of national afforestation programmes and hence to the national forest and
development policy of the country.
(2) A number of prerequisites, e.g. availability of land and a substantial scale
of afforestation, are prerequisites to justify the initiation of a tree improvement pro-
gra
(3) If a tree improvement programme is to be started and is to have a good chance
of success, it must have clearly defined objectives and priorities, and some assurance of
continuing provision of adequate resources to achieve them. As far as possible objectives
should be unequivocal and quantitative.
(4) The tree breeder must examine objectives closely in the light of biological,
human or financial constraints and, if necessary, he must convince policy makers to change
the objectives. Only after realistic objectives have been set and accepted can the tree
breeder start to plan his strategies and programme as described in another lecture.
(5) Tree improvement interacts with other aspects of research and management,
e.g. site assessment and preparation, spacing and thinning. So tree improvement research
must be closely integrated with other sectors of the national forest research programme.
(6) Tree improvement programmes should be flexible and provision should be made
for regular revision. National economic priorities change and may call for changes in
tree improvement objectives. Equally, successes and failures in tree improvement
research may suggest changes in forest management and afforestation policy (change of
species, ability to afforest new sites hitherto considered uneconomic to plant).
Communication between policy and research must be two way.
BIBLIOGRAPHY
Hughes, J.F. & Willan, R.L.
(1976)
Keiding, H.
(1974)
Zobel, B.
(1969)
Polltica, planificaci6n y objetivos. En; Manual sobre
Investigaciones de Especies y Procedencias con Referencia
Especial a los Tr6picos. Tropical For. Paper No. 10 &
lOa. Oxford, UK.
Tree improvement in relation to national forest policy.
In report of FAO/DANIDA Training Course on Forest Tree
Improvement, Kenya. FAO Rome.
A tree improvement program for a developing country and
the effects of tree improvement on forest management.
_In FAO - North Carolina State Forest Tree Improvement
Training Centre Lecture Notes, Raleigh, USA.
- 15 -
ELEMENTS AND PRINCIPLES OF GENETICS I/
W.H. Barrett
Fiplasto S.A., Buenos Aires, Argentina
CONTENTS
Definitions
Chromosome structure
Gene structure
Mutations
Non-genie inheritance 17
Genotype and phenotype 17
Cell division and new gene combinations 18
Fertilization and seed development 18
Endogamy 18
Hybrid vigour 21
Bibliography 21
DEFINITIONS
Genetics is the study of the similarities and differences between individuals, or,
in other words, of variation and inheritance in living things and consequently their
evolutionary process. A distinction may be made between conventional or Mendelian (or
transmission) genetics, and population genetics (the specific effects of genes on living
things). Genetic improvement is the application of the science of genetics for the good
of humanity.
This article describes some basic concepts of genetics, which are applicable both
to animals and plants. Some examples are given of their application to tree species.
CHROMOSOME STRUCTURE
All living cells are made up of an outer cell wall, cytoplasm and nucleus. Within
the nucleus are the chromosomes, which are the carriers of genetic information and are
responsible for transmitting this information to the other cells. They are constant in
number for each species. Thus, for example, willow and poplar species have n-19 chromo-
somes, pines n-12, eucalyptus n-11, etc. Cell size varies according to the species,
usually small with angiospers (which are only a few microns long) and larger with
gymnosperms.
I/ This paper is based on:
- Wright, J.W. (1962). Genetics of Forest Tree Improvement. Forestry and Forest
Products studies No. 16, FAO, Fome;
- Wright, J.W. (1976). Introduction to Forest Genetics. Academic Press.
- 16 -
The chromosome is a long thread-like structure consisting of deoxyribonucleic acid
(DNA) and a protein sheath. DNA, the active genetic material, is a long molecule composed
of two spiral strands. Each strand is composed of four organic bases (cytosine, guanine,
adenine and thymine) and attached sugar radicals; the two strands are joined together by
phosphate radicals. This group generates a molecule from a nucleotide. The two strands
are held together more loosely by hydrogen atoms forming molecules of nucleic acid.
The ability of DNA to replicate itself makes it possible for the chromosomes to
transmit genetic information from one generation to the next. This is a consequence of
the double-strand nature of the molecule and of the properties of the four protein bases.
Adenine and thymine (purinea with two rings) are joined together by two hydrogen bonds,
and guanine and cytosine (pyrimidines with one ring) by three hydrogen bonds. For this
reason, guanine will always be linked with cytosine, and thymine with adenine. At the
time of chromosome division the two strands unravel , and are replicated, and pair off in
such a way that each protein base is linked with its complementary base (adenine with
thymine and guanine with cytosine). To give an idea of dimensions the distance covered
by ten nucleotides on one strand is 34 8. The bases are arranged linearly along a single
strand in groups of three, and, as there are four bases, there are 64 combinations. In
other words, they are like a four-letter alphabet with 64 three-letter words, which are
organized in special sequences or paragraphs which direct growth processes. The sequences
responsible for the formation of some amino-acids and for the activity of certain enzymes
governing protein synthesis are known for some one-celled plants. The molecular weight of
the DNA molecule is very heavy. It has been estimated that a single pine cell contains
some 50 000 000 000 nucleotide pairs. If each nucleotide is considered a letter of the
alphabet, some 2 500 000 pages could be written.
The control of growth processes of a cell is accomplished by a substance similar to
DNA called ribonucleic acid (RNA), which differs structurally from DNA inasmuch as it is
single-stranded. RNA sugars have one more oxygen atom (ribose) and uracil replaces thymine.
RNA performs the function of carrying messages between the DNA and the other parts of the
cell, and regulates and participates in the synthesis of amino-acids, and in the synthesis
of proteins from amino-acids.
GENE STRUCTURE
From a structural point of view, a gene is defined as a sequence of triplets along
a DNA molecule. However, so far it has not been possible to isolate and study the struc-
ture of a gene. A gene can also be defined as that part of a chromosome responsible for
the development of a particular trait. Thus we may speak of genes for rapid growth,
resistance to cold, leaf, length, etc. The gene is considered as the ultimate hereditary
unit, although we know it is of large molecular size. The number of genes in an indivi-
dual cell is unknown, although there are estimates for some plants, such as, for example,
the Pinus banks! ana, in which there are said to be 13 000 000 genes. In some agricultural
crops it has been found that individuals of the same species can differ by 500 genes. For
an improvement programme, a few dozen genes could be used. The genes can have large or
small effects. For example, growth rate is usually governed by many genes with small
individual effects. On the other hand, the colour of eyes in human beings is governed by
a single pair of genes with relatively large effects. Some genes may act independently;
some, however, may act only when others are present.
Genes are arranged linearly on the chromosomes. For this reason, the genes of one
chromosome belong to the same linkage group. However, this linkage is not perfect. At
meiosis, when homologous chromosomes pair, they break and exchange parts. When this
happens, it is said that there has been a "cross-over 11 , giving rise to new genetic combi-
nations. Usually there are one or more cross-overs per chromosome. Logically, the
probability that two linked genes will separate from each other is proportionate to the
distance between them on the same chromosome. This fact is used to prepare chromosome
maps, where the distance between genes is defined in terms of frequency of new combina-
tions; one cross-over unit is equal to 1 percent new combinations. Despite the fact that
gene linkage has not yet been measured in trees, a knowledge of it is Important for tree
breeding. For example, in pines there are 12 pairs of chromosomes and therefore 12
linkage groups. When selecting for any one character, there are apt to be changes in the
frequency of other, closely linked genes.
- 17 -
Genes occupying the same locus in a pair of chromosomes are called "alleles" or
"allelomorphs" and are denoted by a single letter. Dominant allelic genes cause a
characteristic to be expressed even when the tree is heterozygous, whereas recessive genes
only have a visible effect if the tree is homozygous for them. This type of dominance is
complete. It is partial or incomplete when the characteristic is intermediary in expres-
sion. This is what happens with red flowers (AA), white flowers and pind flowers (Aa).
When dominance occurs between non-allelic genes it is called epistatsis.
Genes are called additive when they enhance each other's effects in a cumulative
manner. They are genes with small effects, that govern a single trait.
MUTATIONS
An error in the replicating process of chromosomes produces changes which are
called mutations. These usually occur at the level of the genes; they can cause a gross
change in chromosomes, either in number, structure, inversion, addition or reduction in
size or replication, etc. This produces large effects, usually abnormalities most of
which are harmful.
The frequency of gene mutations is estimated at between 1 in 10 000 and 1 in
100 000 000. The majority of these mutations are harmful and recessive. Despite that,
it is a beneficial process, since it is the principal source of variation of living things.
These changes, because they are recessive, are incapable of expressing themselves. How-
ever, genes which may be detrimental in one environment may be beneficial in another en-
vironment. Gene mutation can be increased artificially through the use of X-rays,
chemical substances, etc. These chemical substances can act directly on the DNA,
producing chemical changes which, during replication, produce mutant offspring. This
happens with nitrous acid or ethanosulphonate ethyl. On the other hand, other substances,
like 5-bromo uracil, only act during the DNA synthesis. By analogy of bases, it resembles
and replaces thymine in replication. During replication this base can occasionally be
coupled with guanine instead of adenine, producing a different triplet.
NON-GENIC INHERITANCE
Cytoplasmic or maternal inheritance has not yet been observed in trees, although
it has in herbaceous plants. In maize a plant was found which did not produce masculine
flowers and was therefore sterile, due to a cytoplasmic factor. This trait was considered
very useful for the production of hybrid maize. Later it was discovered that the same
clone was very susceptible to disease, and it was therefore rejected.
There have been cases of paternal inheritance with a Japanese conifer, Cryptomeria
japonica.
GENOTYPE AND PHENOTYPE
Genotype is defined as the genetic constitution of an individual or groups of
individuals with a similar gene make-up as compared with specific genes.
Phenotype is the external appearance, partially controlled by the genotype. How-
ever, there are recessive genes that are expressed due to the presence of dominant genes,
though they do exist in the genotype. There are also genes with small effects, or modi-
fying genes, that are not manifest in the phenotype. Phenotype is also partly determined
by the environment. A genotype can contain many genes for rapid growth, without this
being apparent, because it is on poor soil or in an adverse climate. Or it may have
genes susceptible to a disease, and appear resistant because the disease is not found
at that site. However, some characteristics of the genotype may be deduced from a careful
study of the phenotype. Of course, much more information will be obtained from studying
a tree's offspring and the relationships between the characteristics of the parents and
those of the offspring.
- 18 -
CELL DIVISION AND NEW GENE COMBINATIONS
The cell division that occurs in the cambium, root, tips, leaves and other growth
mer is terns, is called mitosis* The regular constitution of the cells (2n-diploid) is
maintained by replicating each chromosome constituting two Identical new chromosomes.
Here there is no difference between the chromosomes, apart from possible mutations.
Conversely, in meiosis, there is a reduction In the chromosome number from 2n in
vegetative cells (diploids) to In (haploid) in the gametes. During meiosis, the two sets
of homologous chromosomes come together in the centre of the mother cell, and pair there.
This is where cross-overs occur. Segments of each pair of chromosomes are Interchanged,
producing new gene combinations. Meiosis is a two-stage process, in which there are two
successive nuclear divisions that end by forming four gametes.
FERTILIZATION AND SEED DEVELOPMENT
In fertilization, the male gamete (n) comes together with the female gamete (n),
to produce the egg cell (2n) which through successive divisions forms the embryo, the
seed and later the tree. During this process, chromosomes from the male gamete are in-
troduced into the nucleus of the egg cell. Meiosis and fertilization are the mechanisms
whereby genetic variability makes it possible for individuals to segregate and re combine
and produce offspring genetically different from the parents.
Anglosperms are characterized by dual fertilization; i.e. one nucleus of the male
gamete fertilizes two polar nuclei making up the endosperm. As a result, various geneti-
cally different tissues may be found in the seeds of the angiosperms.
Tissues Number of chromosomes Origin
Seed Coat 2n maternal
Embryo 2n maternal and paternal
Endosperm 3n 2/3 maternal 1/3 paternal
In gymnosperms, however, the endosperm is haploid and of maternal origin.
ENDOGAMY
Endogamy of inbreeding is the crossing of related individuals with loss of hetero-
zygosity. The extreme case is "self ing", or the crossing of a tree with Itself. There
are all gradations from this extreme, for example, back-crossing with one of the parents,
crossing between siblings, cousins, etc., or among members of small Isolated populations.
Wild populations (forests where there has been no human activity) usually carry a
large "genetic load" with a considerable number of recessive genes. Since the Individuals
cross with each other, and not all of them have the same recesslves, these recessive
traits do not appear in their offspring. However, if the tree is selfed, for each hetero-
zygous character 25 percent of individuals among the offspring will show the effect of
this recessive gene. Aa x Aa - 25% AA + 50% Aa + 25% aa. It can happen that the effects
of these harmful recessive genes are small and not visibly obvious, but if the selfed
parent contains 100 recessive genes their cumulative effect on the offspring can be
important. Experiments made with Pinus taeda in the United States showed a loss of
vigour among offspring of selfed trees as compared with open-pollinated offspring of the
same trees. With other forest species, selfed seedlings have often grown less than 50
percent as fast as normal seedlings.
However, self ing does not always bring with it a loss of vigour. Some autogamous
plants, such as tomato and wheat, have flower structures that promote self ing. These
plants may have contained, or acquired through mutations, harmful genes, which have been
eliminated by thousands of generations of self ing. As far as may be discovered, loss of
vigour is not due to self ing or to inbreeding per se, but to accumulations of harmful
recessive genes* A tree free of these recessive genes could be selfed without loss of
vigour.
- 19 -
Allogatnous plants can have mechanisms that inhibit self-pollination, a factor that
leads to a reduction in the quantity of seed. One of these mechanisms is the existence of
genes for self-sterility, like the multiple alleles 8^, 82, 83, 84 etc., where neither $1
nor 82 pollen can germinate a style with the genotype Si 82 etc. There are 40 known
alleles in red clover. However, it can happen that the pollen fertilizes the egg, but the
recessive genes act against the embryo, thus producing a sterile seed.
There is a wealth of literature on the effects of self ing on tree species.
J.W. Wright quotes examples of loss of vigour and reduction in seed fertility in eucalyp-
tus, pine, elms, larches, etc., and others with variable results, ranging from vigorous
to weak offspring and from high to low seed fertility.
This extreme endogamy selfing produces individuals or lines scarcely or not at all
adapted to the surrounding environment. Without reaching this extreme, when the crossing
occurs between a few trees, some random gene fixations take place producing uniform indi-
viduals little adapted to the environment.
To sum up, the general effects of endogamy, fixation of characters, decline in
fertility, size and vigour, are due to an increase in homozygosity.
Measuring endogamy: In the special case of selfing, heterozygosity is lost at the
rate of 50 percent in each generation; F is the endogamy coefficient used to measure the
loss of heterozygosity or increase of homozygosity. The symbol AF is used to express
the per generation change in amount of inbreeding.
A? - 1
2N
N being number of individuals
In this case N is the total number of individuals where each tree is used both as 'male
and female parent. When they are different, the formula to be used is:
+ N (g)
8
To calculate inbreeding over a number of generations, it is better to work with the
heterozygosity coefficient, H 1 - F.
self fertilization
double first cousins
octuple third cousins
generations
Figure 1. Percentage of homozygosity in successive generations under inbreeding
various systems (from 8. Wright, 1921. Systems of Mating, Genetics 6: 167),
To calculate Hn, n being the number of generations, Hn - (H) n ; and H the
heterozygosity per generation. As numerical example, assume that N 5 andA F 1
_
H (per generation) - 9 ; H 2 - (9)2 - 0.810; H 3 -
10 (10) (10)
0.729
n - 10
n 100
0.651
0.999
0.401
0.994
0.182
0.86
0.095
0.63
0.0489
0.39
0.0198
0.18
- 20 -
After three generations, 72.9 percent of the original heterozygosity remains in the
population, and the inbreeding rate is:
F 3 - 1 - H 3 - 1 - 0.729 - 0.271
Inbreeding coefficients of populations maintained as sites of N - 50 to 250 for
10 or 100 generations.
Population Coefficient F
N
5
10
25
50
100
250
(extract from J.W. Wright, 1976)
In tree breeding it is assumed that the loss of vigour through unbreedlng is pro-
portionate to the loss of heterozygosity as estimated by the coefficient F. In annual
plants, on the contrary, it is easy to take five individuals and reproduce them over
successive generations and measure the results.
There are mechanisms in plants that encourage or prevent autogamy. Annual plants
usually have perfect hermaphrodite flowers and have developed mechanisms suitable for
self ing. This is not generally the way with trees, except (according to J.W. Wright,
1976) for tropical trees constituting mixed forests with hundreds of species per hectare,
where the possibility of crossing is remote and there is therefore a high probability of
self Ing or at least a high percentage of endogamy. There must therefore be mechanisms
to prevent this damaging the species. Unfortunately we have no experimental data on
this subject.
In temperate or cold climates, however, where tree species tend to grow in stands
which are pure, or composed of a few species only, there are mechanisms inhibiting self-
ing. In these cases a greater variability is necessary since there are great fluctuations
in rainfall, temperature, exposure to fire, glaclatlon, etc. One of these mechanisms is
the dioecious character of the plants, e.g. male and female flowers are produced on
different trees. This happens with poplars, willows, junipers, Araucaria araucana,
A. angustlfolia, etc. Monoeclsm also performs this function, e.g. male and female
flowers appear on the same tree, as with oaks, Aesculus , and most conifers. In the
latter group the masculine flowers are produced on the lower branches, and the female
flowers on the higher ones. As pollination is wind-borne, this stratification inhibits
self Ing to a large extent, since the male pollen would have to rise.
There are mechanisms like dichogamy, whether in hermaphrodite or In diclinous
flowers, where the male flower Is receptive at a different time than the female; for
example, the hazel tree with diclinous flowers (protandry) and the tulip tree
(Liriodendron) with hermaphrodite flowers (protogeny) .
A simple method has been developed by forest geneticists to determine the per-
centage of self ing in natural populations of different forest species . It consists in
counting the number of abnormal seedlings from open pollinated seed, in the nurseries
and multiplying the figure by 5. In various pine species between 2 and 7 percent of
self ing has been recorded.
"Self ing plus later crossing" has been used to improve herbaceous plants, but it
lacks practical possibilities when applied to trees.
The random fixation of genes in small populations is called genetic drift. If
inbreeding is continued over a long geological period, the result Is a uniform population
which hae developed non-adaptive traits. The isolation of small pine populations In the
mountains of Mexico has produced a great diversity of forms, which is probably due to
random gene fixation. Although this is more frequent in mountain areas, it has also been
- 21 -
observed In existing islands, or regions that were under water in different geological
periods. This is the case of Pinus elliottii var. densa which has highly differentiated
morphological and adaptive variations, irrespective of distance, in the absence of
clonal gradients. One may observe on one island key, for instance, a population which
has fascicules with two relatively short leaves and, on a neighbouring key, fascicules
with three long leaves.
J.W. Wright (1976) observed similar examples in species of Picea, noting that many
of these populations which have not adapted to the environment in which they live are on
the point of extinction.
HYBRID VIGOUR
Hybrid vigour or heterosis refers to the exceptional yield of a hybrid when com-
pared with its parents. To explain this phenomenon, four explanations have been suggested,
which are: dominance, over dominance, additive and hybrid habitat hypotheses.
According to the dominance hypothesis, hybrid vigour is due to the absence of
depression caused by endogamy. The homozygosity produced by inbreeding leads to the
appearance of recessive genes which are harmful to the plant, producing a depression
which would otherwise be "covered" by dominant genes from heterozygous individuals.
The overdominance hypothesis maintains that the heterozygous gene combination
produces effects impossible to achieve with homozygous genes.
The additive hypothesis holds that the trait is governed by various genes which
hybrids accumulate, the effect being thereby greater.
In the "hybrid habitat" hypothesis, vigour derives from a "hybrid" environment,
where the parents show less capacity for adaptation than the hybrid.
If the first hypothesis, of dominance, is true, it should be possible to select
against recessive genes in subsequent generations and thus fix hybrid vigour. If it is
due to overdominance, vigour will be greater in the first generation and decrease in
successive ones. If it is due to additive effects, it should be possible to select the
best genes governing the components of this vigour, and to obtain a greater effect in
successive generations. Lastly, if the hybrid habitat hypothesis is true, hybrid vigour
will be equal in the generations following the F]_ generation.
A case of extensive utilization of hybrid vigour was the crossing of Pinus rigida
with . taeda in South Korea, where performance was excellent in F^ and even better in F2-
The breeding method depends on the type of inheritance of the trait to be bred
for. If greater productivity is due to dominance or additive effects, selection should
exclude self ing. If, on the other hand, vigour is due to over dominance "self ing and
later crossing" is the best method.
BIBLIOGRAPHY
Beadle G.W. and M. Beadle, 1971. Introducci6n a la nueva Genetica (translated from
English). Editorial Universitaria de Buenos Aires, Argentina, 282 p.
Herkowitz I.H., 1965. Genetics (2nd ed.), Little, Brown and Co., Boston, 554 p.
(Spanish translation, GenStica D.E.C.S.A. Mexico, 765 p. 1970).
Lacadena J.R., 1976. Gengtica (2nd ed.), A.G.E.S.A., Madrid, 972 p.
SRB, A.M., R.D. Owen and R.S. Edgar, 1965. General Genetics, Freeman and Co., San
Francisco (Spanish translation, Ed. Omega, Barcelona, 632 p. 1974).
Strickberger M.W., 1976, Genetics (2nd ed.), Macmillan Publ. Co. Inc., N.Y. 914 p.
(Spanish translation, 2nd ed. Omega, Barcelona 1978).
- 22 -
Wright, J.W. 1976. Introduction to forest genetics. Academic Press, 463 p.
Wright, J.W. 1962. Genetics of Forest Tree Improvement. Forestry and Forest Products
Studies, No. 16, FAO, Rome, 399 p.
Participants in class.
- 23 -
PRINCIPLES AND STRATEGIES FOR THE IMPROVED USE OF FOREST GENETIC RESOURCES
C. Palmberg
Forest Resources Division
Forestry Department
FAO
CONTENTS
Page
Introduction 24
Principles of conservation and utilization of genetic resources of forest trees . 24
Exploration 25
Collection for Evaluation 25
Evaluation 25
Conservation 26
In situ 26
Ex situ 27
Collection for conservation ex situ 28
Storage as seed 28
Conservation stands 28
Dissemination of information 30
Utilization 30
The need for international action 30
Concluding remarks 31
References 31
Annex 1. Genetic Resources of Forest Trees: Phases and Operations 35
Annex 2. Genetic Resources of Forest Trees: Duration of necessary phases;
hypothetical example for a tropical pine 36
Annex 3. Progress in the Conservation and Utilization of Forest Tree
Genetic Resources 37
- 24 -
INTRODUCTION
Increase in the world's population, together with higher standards of living, result
in continuous pressure to transfer areas previously under forest to agricultural or other
uses (Villan 1973). The resulting large-scale disappearance of natural forests is leading
to an accelerated loss of valuable or potentially valuable germplasm. This loss 5s of
particular concern in those areas in which botanical and genecological exploration has not
been systematically carried out, and in which species composition as well as inter- and
intra-specif ic variation therefore are not known sufficiently well to enable timely and
adequate conservation measures.
In addition to the fact that large areas of forests are being partially or fully
destroyed, areas destined to remain as forest are often being brought under more intensive
forms of management which may endanger certain species and change the genetic composition
of others (Kemp et. al. 1976). Even where the central part of the range of a species is
unnaffected, particular sub-population or provenances at the limits of the species range
may be in critical danger; it is often these marginal or isolated populations which have
developed, through natural selection, specific characteristics such as tolerance to drought
or other adverse environmental conditions, and which may be of great potential value for
sites with a similar selection pressure.
The continuous pressure for land mentioned above coupled with an increasing demand
for wood and wood products has tended to shift the emphasis from utilization of the often
complex natural forests to plantations of species relatively easy to manage and capable of
producing large quantities of wood per unit area (Willan 1973). Although the creation of
plantations to a certain extent will relieve the pressure on the natural forests and the
genetic material they contain, their establishment may raise problems. The use of plantations
gives the forester an opportunity to exercise much stricter control not only over" site
characteristics but also over the genetic quality of his forests (Willan and Palmberg 1974).
This leads to a development from the use of 'wild 1 to the use of more 'advanced 1 populations,
in which gene frequencies have been changed to meet specific requirements. In these new
populations, selected and bred for uniformity, high yield and other short-term objectives,
the genetic base is often narrowed down to very low levels by restricting the genepool from
which parental material is drawn and by subsequently rejecting, through selection in given
conditions, a great proportion of the original population. While adaptation of these new
populations to specific plantation conditions is increased, their genetic flexibility and
their potential for future adaptive change to meet often unforseen or unforseeable environ-
mental changes such as a shift in the average site quality of plantations, the emergence of
new or genetically adapted pests and diseases, or increased levels of industrial pollution,
is gradually decreased. The narrowing of the genetic base in populations used for the
production of seed for future plantations does not, as such, necessarily have negative effects,
as long as the genetic diversity of the species and the provenances is safeguarded through
conservation measures in situ or through the establishment of genetic reserves, conservation
stands and/or widely-based breeding or base populations, from which material to meet new
requirements can be drawn.
PRINCIPLES OF CONSERVATION AND UTILIZATION OF GENETIC RESOURCES OF FOREST TREES
Conceptually, the issues of genetic conservation are similar whether dealing with
annuals or long-lived trees, domesticates or wild plants; needs, strategies and methods
differ in detail, but not in principle (Frankel 1978).
The exact strategy of conservation depends on the nature of the material and the
objective and scope of conservation. The nature of the material is defined by the length of
the life cycle, the mode of reproduction, and the ecological status of the individuals (wild,
domesticated); the objective could be research, static or evolutionary conservation (see
below), selection and breeding, etc.; the scope refers to timescale and area considered.
(Frankel 1970).
Activities commonly recognized as essential steps to maintainance of genetic diversity
within individual species and to the fuller use of existing genetic resources are (i) explo-
ration; (ii) collection; (iii) evaluation; (iv) conservation and (v) utilization. (FAO,
I975a - See Annexes 1 and 2).
- 25 -
EXPLORATION
Efficient use of existing genetic resources can only be achieved if sufficient infor-
mation is available on their extent, structure and composition (Brazier <et. aU 1976; Sneep &
Hendrikson 1979; Lamprey 1975). For a large number of tree species, especially for species
growing in the tropics, there is a great lack of knowledge on the ecology and biology, as well
as on their potential as plantation species and the potential use of non-wood products derived
from them. Even for species of proven value, their variation throughout their natural range
has often been explored inadequately. This lack of information is well illustrated by the
fact that every one of the tropical/sub-tropical forest tree species explored and collected
during the present decade in a programme coordinated by FAO has, as a consequence of the
exploration, been found to be in danger of depletion, extinction or contamination of its
genetic resources in at least parts of its natural range; where the genepool is not thought
to be in danger of extinction, the population has often been so heavily reduced that supplies
of seed are very limited and may deplete further in future (Keiding and Kemp 1978).
For practical purposes, the field activities in the fundamental step of exploration
can be divided into (i) botanical and (ii) genecological exploration. Botanical exploration
includes the correct taxonomic identification of species and knowledge of the limits of their
distribution, with particular reference to isolated occurrances. For some forest tree
species adequate information has been available well before the start of genecological ex-
ploration, for others it may be necessary to combine the two operations together. Botanical
exploration logically leads to species trials.
Through genecological exploration patterns of ecological and phenotypic variation
within the natural range of species are studied, leading to provenance seed collections
and provenance evaluation (FAO 1975).
COLLECTION FOR EVALUATION
Collection for evaluation consists in collecting relatively small samples of seed from
a relatively large number of seed sources, covering the whole natural range of the species.
In the initial stage, collections thus comprise range-wide sampling on a fairly coarse grid;
in some cases a second stage, sampling a limited part of the range on a finer network, may
be called for after the results of first stage provenance trials are available. In these
second-stage collections seed is often kept separate by mother trees to enable the evaluation
of genetic variation within, as well as between, provenances. The genepool s included in
the collections may be indigenous, or introduced. In forestry, these latter so-called 'land
races 1 (i.e. exotic plantations which to various degrees have adapted to local conditions
as a response to natural and sometimes artificial selection), are potentially of great
importance as sources of seed, and should be included in the collections (FAO 1975a,
Turnbull 1978.
To determine the number and location of populations to be sampled, environmental
gradients are generally followed; sampling within each population could be done at random
or selectively. Although the latter option is often used in within population sampling
it should be remembered that phenotypic superiority is no guarantee of genetic superiority,
specially in cases in which the past history of the population is unknown (Barner 1974;
Bennett 1970).
EVALUATION
Collection of range-wide samples should be followed by the establishment of provenance
trials aimed at revealing potentially useful variability, degree of adaptation to a range
of environmental conditions, and economic or social value of the species/provenances under
test. The evaluation should be carried out on as many potential sites as possible and,
whenever feasible, be centrally coordinated.
- 26 -
CONSERVATION
The development of the concept of genetic conservation in the 1950s was mainly due
to the realization that the primitive cultivars of traditional agriculture were rapidly
disappearing, and the genetic diversity accumulated in them over many centuries substituted
by varieties selected and bred to meet short-term needs. The importance of maintaining
genetic diversity and genepools from which new genes can be introduced into existing
varieties in order to improve adaptation, yield and resistance to diseases and adverse
conditions, has clearly been brought into focus by some major outbreaks of diseases, espe-
cially in food crops (Frankel 1978, Sneep and Hendriksen 1979). Not only should variability
within species of known value be conserved, but maximum diversity should also be maintained
on a species level, including as yet unknown and untested material, thus keeping future
options open (Whit mo re I975a).
Conservation, used in its proper sense, embraces both preservation and utilization;
conservation is, in fact, an aspect of resource management which ensures that utilization
of the resource is sustainable, at the same time safeguarding genetic diversity essential
for its maintenance.
A compromise between biological, technical, economic and administrative factors is
often unavoidable when choosing long-term strategies for the conservation and utilization
of genetic resources. The final objective must be to choose methods which minimize losses
and maximize gains in terms of usefulness, knowledge and integrity (Frankel 1970a),
Current problems in genetic conservation are often so severe that it is tempting
to concentrate on them alone. However, strategies for action should include preventive
measures through the inclusion of conservation aspect in long-term planning at the policy-
making, the organizational as well as the technical level (Anon 1980).
The main strategies for conservation have been outlined as follows (Burley and
Styles 1976):
1. Conservation of ecosystems. Such conservation of carefully selected areas of
adequate size and with suitable management policies would preserve not only
forest trees but also other elements of the ecosystem (plants, mammals, birds,
etc.), as well as other potentially valuable products such as extrativas,
fruits, etc.
2. Preservation of rare species and species threatened with extinction. This
goal could often be achieved under general ecosystem conservation, if carried
out competently.
3. Prevention of "genetic erosion 11 , i.e. depletion of genetic variability. In
this field it is not enough to conserve a species as such; we must make certain
to conserve a broad spectrum of genetic variability to act as a reserve for
present and future needs (adequate seed sources, wide genetic variability as
a base for tree breeding , etc.). Such material can be conserved in in s i t u
reserves, or can be sampled by means of seed, pollen or vegetative material,
in such a way that the conservation of the major part of the genetic variability
is insured. Seed, pollen or other material can either be stored as such, or
be used for the establishment of conservation stands ex situ.
IN SITU CONSERVATION
Conservation jln situ, i.e. conservation of species /provenances as part of a viable,
existing ecosystem, is generally the most desirable method of conserving forest genetic
resources, provided that the area can be given full protection and provided the genetic
material conserved is made available for use both within and outside the country of origin
(FAO 1975; Whitmore 1975a, b; Lamprey 1975; IUCN 1978). For many species, e.g. for a vast
number of rainforest species which are not fast-growing pioneers, which occur as individuals
rather than stands, and for which knowledge on ecology and genetics is scarce or lacking,
in situ is the only method of conservation available to us at our present level of
knowledge (Kempt 1978).
- 27 -
In situ conservation of forest genetic resources will often for practical reasons
have to be combined with other environmental, scientific or socio-economic purposes, this
generally means that a compromise must be made between various objectives of the reserve.
Genepool conservation frequently deals with genetic differences which cannot be
directly identified but only surmised. It is concerned with population samples, possibly'
along latitudinal or altitudinal transects, often over extensive areas, to include a spec-
trum of ecological variability to provide a corresponding spectrum of genetic variability.
The efficiency of ecosystem conservation (e.g. biosphere reserves, national parks) to
adequately meet the needs of genepool conservation is closely related to the size, number,
distribution and location of these reserves.
There is a general consensus that the conservation of representative samples of
most ecosystems will require an area within the range of 100 to 1 000 ha, the exact size
depending on the heterogenity of the area as well as on its species composition (Ashton
1976). If conservation of genetic resources is a major objective of ecosystem conserva-
tion, the inclusion within the reserve of the minimum number of interbreeding individuals
needed for a viable genepool (i.e. a population which is able to retain its self-renewing
capacity), must be considered rather than the total area of the reserve per s (Roche 1975),
Considering genetic resources at a species level , Ashton (1976) working with rain-
forest species in Borneo made a theoretical estimation of the area of forest needed for
conservation, arbitrarily assuming that 200 mature individuals will form a viable popula-
tion; by this criterion an area of at least 2 000 ha of unmodified virgin forest would
have been needed to conserve the tree species of two areas examined, while only 60% of the
species would have been safeguarded in 1 000 ha. At the intra-specific level Dyson (1974),
studying numbers of individuals quoted for effective breeding populations in animals,
estimates that 200 individuals will constitute a minimum "safe" population for forest
trees for maintaining genetic variability, provided that sampling is done from at least
three parts of the range. Up to 25 000 individuals are recommended by Marshal (quoted
in Kemp 1978) 'for maintaining a given level of heterozygosity in populations in "a dif-
fusely outbreeding tree species'. However, as discussed below, the conservation of a
mythic heterozygosity (Namkoong 1979a) by conserving specific genotypes is neither
desirable nor possible.
Theories as to the relative merits of one large as opposed to several smaller,
separate reserves, have been extensively discussed. The answer will depend on the exact
objectives of conservation, the amount of inter- and intra-specific variation to be con-
sidered, and the distribution of gene frequencies. From a purely managerial point of
view, one or very few large areas would be preferable, as a large number of scattered re-
serves are difficult to manage and protect. However, especially in the case of areas with
a complex species composition, and in the case of conserving the intra-specific variation
of widely distributed species, a series of reserves, strategically placed to sample the
full range of ecological variation is called for. It is particularly important to include
extreme environments and marginal populations, in which selection effects may have created
varieties or ecotypes of particular potential value, and in which gene frequencies may
differ from those in the main population, giving us bigger chances to capture 'rare genes'
(Namkoong 1979a, b).
EX SITU CONSERVATION
Although conservation In situ in theory may be the most effective strategy it may,
in reality, face enormous difficulties which often are social, political or financial
rather than technical in nature (Sastrapradja et al. 1978; Kemp et. al. 1976). The alter-
native way of conservation is ex situ. Ex situ conservation is especially useful when
dealing with certain species or genera with a combination of biological characteristics
which make them amenable to this approach; through knowledge of the breeding system and
biology of the species, as well as of the methodology for growing them in plantation con-
dition and/or storing their seed, are prerequisites for using this strategy. Many of the
species which during recent years have attracted the attention of foresters for their use
in high-yielding plantations fall into this category.
Sometimes, especially in the case of economically valuable plantation species ex-
tensive genetic modification of native stands caused by man may occur (Libby 1978). This
situation arises if non-native populations of a species are used for seed procurement and
the subsequent establishment of plantations close to existing native stands. Clouds of
- 28 -
pollen from the plantations will repeatedly be dispersed over the native stands, resulting
in offspring which will be increasingly contaminated by genes from the foreign populations,
gradually leading to a loss of the original genepool. In these cases, in situ conservation
will not be feasible and the only way to conserve the original population will be by ex
situ conservation.
COLLECTION FOR EX SITU CONSERVATION
Where the exploration phase has shown that certain populations are endangered but
that conservation ^n situ is not feasible, early collection is necessary of substantial
quantities of seed or other propagating material of endangered provenances either for tem-
porary storage or for immediate establishment of ex situ conservation stands on new sites
(FAO 1975). Sampling for variation (i.e. random rather than selective sampling) is
essential for conserving the integrity of allele frequencies (Frankel 1970b).
Methods of sampling for gene conservation purposes and theoretical number of indivi-
duals needed to maintain intraspecific allelic variation are discussed by Namkoong (1979a),
who gives probability levels for losing specific alleles occurring at given frequencies,
using different sampling intensities within and among populations.
It is not possible to lay down general rules and guidelines for sampling, as many
factors, both interrelated and independent, affect the intraspecific variation we are
trying to capture (heterogenity and size of the natural range, ecology, breeding system
and population structure of the species, etc.). However, as no system of sampling and
collection as likely to achieve the objective of saving all allelic combinations present
in a species, sampling is generally aimed at saving as many of the existing alleles as
possible for future recombination and use (Namkoong I979a). We should thus aim at conserving
and evaluating genes rather than genotypes.
STORAGE AS SEED OR OTHER REPRODUCTIVE MATERIALS
In addition to being a means of conservation per se, storage of seed is often an
essential link between collection and later field operations. Meticulous handling of the
seed during all phases of the work is essential. For many species particularly in the
tropics, there is inadequate knowledge of practical methods of both long-term storage, and
additional research is urgently needed.
Conservation of forest trees is generally done as conservation stands jln situ or
x situ, rather than as seed, as is often the case in agricultural crop species. This
difference in approach is mainly due to practical difficulties: plant gene banks should
regenerate their seed collections whenever the viability falls by, at most, fifteen percent
below the initial value at which the seed was stored (Wang 1976; IBPGR 1976). With the long
vegetative period that most tree species have before they produce viable seed, seed regenera-
tion by growing a crop and re-collecting will be a prolonged and expensive procedure; in
addition, natural selection during this prolonged period is likely to have greater effects
on genetic composition than in the case of species which produce seed within a short period
of time after sowing.
With present-day knowledge of the physiology and biochemistry of pollen and tissue
culture, conservation of forest genetic resources in these forms seems unlikely to become
more than a useful supplement to other forms of conservation. Although pollen storage
is a valuable means for short or intermediate-term conservation storage time of pollen,
when using known drying and storing techniques, is generally shorter and less reliable than
that of seed. Likewise, and with the possible exception of vegetatively propagated species*
the conservation of forest genetic resources by means of tissue culture is generally not
considered likely to become of great importance (Wang 1978; Frankel 1978).
EX SITU CONSERVATION STANDS
Conservation stands ejc situ are expensive to establish and to maintain, and therefore
they are normally confined to species of proven or evident potential value (FAO 1975).
Danger of extinct ion, economic potential and difficulty in reprocurement of seed should be the
main criteria in forming priority list of species and provenances for jnc situ tonservation.-
I
- 29 -
Guldager (1978) lists four conservation objectives which may be met by the establishment
of ex situ conservation stands:
(1) Static conservation, in which the genotype frequencies of the original
population are maintained. As discussed above, this method is not practicable for most
forest tree species.
(2) Static conservation, in which the gene (allele) frequencies of the original
population are maintained. No genetic information is lost and any genotype found in the
original population could in principle be reproduced although genotype frequencies in the
ex situ stands might be different than in the original population.
(3) Evolutionary conservation, in which gene frequencies in the stand are allowed
to change according to natural selection pressures.
(4) Selective conservation, in which gene frequencies in the stand are deliberately
changed by man in order to capture characteristics important for plantation economy in a
region, and at the same time eliminate undesirable characteristics. To avoid a decrease
of the genetic potential for future plantation establishment in environments different from
the environment in which this type of stand is established, replication is necessary in
each potential plantation area. In the long term, selective conservation faces the same
problem as met in long term breeding programmes, (i.e. maintenance of genetic variation,
avoidance of inbreeding).
The ex situ conservation stands known to have been established to date fall under
categories IT) and (4). The level of maintenance of genetic integrity in these stands will
be dependent on three main factors (Guldager 1976): (i) sampling in the original popula-
tion; (11) survival and growth ex situ of sampled genotypes (i.e. adaptation to new selection
pressures); (iii) mating patterns ejc situ between the sampled genotypes.
(i) Sampling for conservation has been discussed above* Even the most rigorous
efforts to maintain original gene frequencies through careful management in a series of
conservation stands ex situ will be of little avail from the point of view of species/
provenance conservation if the gene frequencies have already been considerably changed
during initial sampling. Sampling is thus of critical importance. Long term storage or
maltreatment of seed are other factors that may critically influence gene frequencies even
before the stands are established.
(ii) For most plantation species it is possible to combine the selection of suitable
sites with efficient nursery and plantation techniques so as to ensure almost 100% survival
in the field. Initial competition between genotypes can be minimized by wide spacing. The
question of mechanical versus silvicultural thinning in the stands will depend both on the
ultimate objective of conservation and on practical possibilities. However, if the stands
are established over a wide range of sites in which environmental pressures vary, a great
proportion of the genetic variation is likely to be maintained even if thinning is carried
out in favour of desirable phenotypes. As a compromise, a proportion of the trees to be
left (say, 1%) may be phenotypically selected before carrying out a systematic thinning
for the rest of the stand. This approach will be adapted in the case of the international
conservation stands mentioned below.
(iii) Our possibilities to accurately pass on genetic information from the first
generation ex situ conservation stand to the next depends on mating within the stand
(synchronization of flowering, proportion of actual random mating, etc.), size of population
(influencing genetic drift and inbreeding coefficient), and migration in terms of pollen
contamination. To help overcome these problems, location (near optimum or optimum for
flowering and seed production), size (recommended size is 10-30 ha FAO 1975, 1977) and
isolation (300 m or more between hybridizing species/provenances; FAO 1977), should be
carefully considered. In addition to careful siting, the conservation stands require
meticulous standards of site preparation, planting and tending (FAO 1975). An imperative
condition for their establishment in a region is thus that sufficient technical expertise
as well as organizational stability is available, to ensure a high standard of long tern
management. Interest in the provenances concerned from a plantation point of view, is
likely to increase the benefit as well as the security of the scheme (Guldager 1978).
- 30 -
Detailed recommended prescriptions for the establishment and management of ex situ
conservation stands have been published in Annex 7 of the Report of the 4th Session of the
FAO Panel of Experts on Forest Gene Resources (FAO 1977).
Besides the long-term benefits of conserving species/provenances of known genetic
characteristics, conservation stands have valuable possibilities for short-term utilization,
providing seed and other genetic material for immediate use. In cases where international
finance has been provided, agreements have therefore been drawn up to ensure that the stands
benefit all countries interested in the species/provenances (see FAO 1977, Appendix 7).
DISSEMINATION OF INFORMATION
Another aspect of conservation is the preservation and dissemination of information.
It is important to preserve not only the areas, units, populations or individuals, but also
the relevant data on them. This data must be carefully recorded, safeguarded and dissemi-
nated (Frankel 1970a).
UTILIZATION
Utilization is the ultimate objective of all activities concerned with forest genetic
resources. It comprises both the use of bulk supplies of seed or other propagating material
for large-scale plantation schemes, and the breeding of better adapted and more desirable
genotypes.
As information becomes available from provenance trials as to the most suitable seed
sources, emphasis will gradually switch from evaluation to the utilization of bulk supplies
of seed or sometimes other propagating material of populations found well-adapted to given
conditions. The supply of bulk quantities of propagating material should be the* responsibi-
lity of Government Forest Services or commercial seed merchants, although agreements both
within and between countries on common standards of genetic and physiological quality of
the material are essential (FAO, 1975a).
Individual selection and breeding within locally adapted provenances provide a method
of achieving additional improvement in selected characteristics. In the case of introduced
species, an important interim stage between successful provenance trials and large-scale
afforestation with the best-adapted provenances may be the establishment of one or more
substantial blocks (5 ha or more) of these provenances to act as seed stands and as a basis
for local selection and breeding* The same stands may combine the purpose of ex situ
conservation.
NEED FOR INTERNATIONAL ACTION
When the urgency of conservation and the massive efforts needed are jointly con-
sidered, it becomes evident that conservation of the world's genetic resources requires the
cooperation of all nations.
Although progress in development of improved forest genetic resources will remain
largely dependent on the active efforts of individual countries or research institutes,
these can only be fully effective in an international framework (FAO 1975). Maintenance
of the genetic diversity of species either jln situ or eoc situ may have to be spread over
many environments in a number of countries, collection of seed cannot be confined within
national borders; research coordinated to give information of species/provenance performance
over the maximum range of sites possible will be of great mutual benefit to cooperating
institutes and nations; safety and permanence for irreplaceable collections of genetic
material, either j.n situ or ex situ should be secured for perpetuity by agreements under
international supervision.
Many countries which contain forest genetic resources of great but sometimes un-
explored potential value are at an early stage in their economic development. There is
often a severe shortage of funds and skilled staff in the forestry sector and those
available are, logically, channeled to meet immediate national needs when e.g. drawing up
species priorities. It is therefore highly desirable that international resources should
be made available to help in the development of strategies and to safeguard material
invaluable to many nations.
- 31 -
There is another aspect to conservation, and that is the conservation and dissemina-
tion of information. It is not only important to conserve areas, communities, populations
and individuals, it is also important that information related to them is adequately re-
corded, safeguarded and made available (Frankel 1970).
The best way of ensuring efficient coordination in the wide field of forest genetic
resources is by the adoption of a global programme such as that proposed by the FAO Panel
of Experts on Forest Gene Resources (FAO 1975a). Such a program should ensure the inte-
gration of conservation measures with the equally important activities of exploration,
collection and utilization. At the same time it should improve efficiency through coor-
dinating the efforts not only of the many countries but also of the several international
agencies concerned with genetic resources (Roche 1978).
Progress in the conservation of forest gene resources in the last ten years is out-
lined in Annex 3.
CONCLUDING REMARKS
In the rapidly advancing field of forest genetics we have answered many question
in recent years, but these answers have often led to new and more difficult questions.
We have learned enough techniques to be certain that we can develop new breeds to more
accurately meet present-day needs. We have also come to realize more fully that the ori-
ginal gene pools will be lost unless positive action is taken to conserve them. Now we need
to decide how to organize breeding and gene management strategies that will meet both
immediate needs and long-term requirements (Namkoong 1978).
It should not be difficult to conduct tree improvement programmes which include
short-term and long-term objectives side by side, provided that those responsible for
planning and finance realize that the long-term programme is no less important and no less
deserving of funds than the short-term one, and that the greater the genetic diversity we
can maintain and save now, the wider will be our options for finding suitable genetypes
to meet future needs (Namkoong 1979).
REFERENCES
Anon, 1980. World Conservation Strategy. Living Resource Conservation for Development E,
F, S. IUCN/UNEP/WWP/FAO (en prensa).
Ashton, P.S. 1976. Factors Affecting the Development and Conservation of Tree Genetic
Resources in South-East Asia. En; Tropical Trees: Variation, Breeding and Conservation
(Eds. J. Burley & B.T. Styles). Linnean Society, Oxford, U.K.
Barner, H. 1974. Classification of Sources for Procurement of Forest Reproductive Material.
En; Report of the FAO/DANIDA Training Course on Forest Tree Improvement. Limuru, Kenya,
September-October 1973. FAO/DENTF 112. FAO, Roma.
Bennett, E. 1970. Tactics of Plant Exploration. En; Genetic Resources in Plants- their
Exploration and Conservation (Eds. O.H. Frankel & E. Bennett). IBP Handbook No. 11.
Blackwell Scientific Publications. Oxford and Edinburgh.
Brazier, J.D., Hughes, J.F. & Tabb, C.B. 1976. Exploration of Natural Tropical Forest
Resources and the Need for Genetic and Ecological Conservation. En: Tropical Trees:
Variation, Breeding and Conservation (Eds. J. Burley & B.T. StylesT* Linnean Society,
Oxford, U.K.
Burley, J. & Nikles, D.G. 1972 & 1973a. Selection and Breeding to Improve some Tropical
Conifers. Vols I & II. Based on Papers submitted to a Symposium organized by IUFRO
Working Parties S2.02.08 and S2.03.01, held in Gainesville, Florida, USA in 1971.
Commonwealth Forestry Institute, Oxford, U.K.
Burley, J. & Nikles, D.G. I973b. Tropical Provenance and Progeny Research and International
Cooperation. Based on Papers submitted to a Symposium organized by IUFRO Working Parties
S2.02.08 and S2.03.01, held in Nairobi, Kenya in 1973. Commonwealth Forestry Institute,
Oxford, U.K.
- 32 -
Bur Icy, J. & Styles, B.T. (Eds.) 1976. Tropical Trees: Variation, Breeding and
Conservation. Linnean Society, Oxford.
Cromer, D.A.N. 1976. Report on Consultant Mission on Conservation of Forest Genetic
Resources in selected countries in Asia. FO:MISC/76/27. FAO, Rome.
Dyson, W.G. 1975. A Note on the Conservation of tree species in situ. In: Report of the
Third Session of the FAO Panel of Experts on Forest Gene Resources. FO-FGR/3/Rep. FAO
Rome.
FAO, 1973-79. Forest Genetic Resources Information. Forestry occasional Paper. FAO Rome.
FAO, 1969. Report on the First Session of the FAO Panel of Experts on Forest Gene
Resources. FO:FGR/l/Rep. FAO, Rome
FAO, 1972. Report on the First Session of the FAO Panel of Experts on Forest Gene
Resources. FO:FGR/2/Rep. FAO, Rome.
FAO, 1975a. Proposals for a Global Programme for the improved utilization of forest
genetic resources. Forest Genetic Resources Information No. 4 1975/1 FAO, Rome.
FAO, 1975b. Report on the First Session of the FAO Panel of Experts on Forest Gene
Resources. FO:FGR/3/Rep. FAO, Rome
FAO. 1977. Report on the First Session of the FAO Panel of Experts on Forest Gene
Resources. FO:FGR/4/Rep. FAO, Rome
Frankel, O.H. 1970b. Evaluation and Utilization - Introductory Remarks. In: Genetic
Resources in Plants - their Exploration and Conservation (Eds. O.H. Frankel and E.
Bennet). IBP Handbook no. 11. Blackwell Scientific Publications, Oxford and Edinburgh.
Frankel, O.H. 1970a. Genetic Conservation in Perspective. In; Genetic Resources in
Plants - their Exploration and Conservation (Eds. O.H. Frankel and E. Bennett). IBP
Handbook no. 11. Blackwell Scientific Publications, Oxford and Edinburgh.
Frankel, O.H. 1978. Philosophy and Strategy of Genetic Conservation in Plants. Tercera
Consulta Mundial sobre el Mejoramiento de Arboles Forestales. FC:FTB/77-l/2. Canberra,
Australia.
Franke, C.H. & Bennett, E. (Editors) 1970. Genetic Resources in Plants: their Exploration
and Conservation. IBP Handbook no. 11. Blackwell Scientific Publications, Oxford and
Edinburgh.
Guldager, P. 1978. Ex situ Conservation stands in the tropics. In; The Methodology of
Conservation of Forest Genetic Resources. FO:MISC/75/8. FAO. Rome.
IBPGR, 1976. Report of IBPGR Working Group on Engineering, Design and Cost Aspects of Long-
Term Seed Storage Facilities. International Board for Plant Genetic Resources. FAO, Rome.
IUCN, 1978. Categories, objectives and criteria for protected areas. A Final Report
prepared by Committee on Criteria and Nomenclature Commission on National Parks and
Protected Areas. Morges, Switzerland.
Keiding, H. and Kemp, R.H. 1978. Exploration, collection and investigation of gene
resources: tropical pines and teak. Tercera Consulta Mundial sobre el Mejoramiento de
Arboles Forestales. FO-FTB-77-1/3. Canberra, Australia
Kemp, R.H. , 1976. Report on Consultant Mission on Conservation of Forest Genetic
Resources in selected countries in Africa. FO:MISC/76/26. FAO, Rome.
Kemp, R.H., 1978. Exploration, utilization and conservation of genetic resources. Third
World Consultation on Forest Tree Breeding. FO-FTB.77.1/1.
- 33 -
Kemp, R.H. , Roche, L. & Will an, R.L., 1976. Current activities and problems in the
exploration and conservation of tropical forest gene resources. In: Tropical Trees:
Variation, Breeding and Conservation (Ed. J. Bur ley & B.T. StylesTT Linnean Society,
Oxford, U.K.
Kemp, R.H. & Whltmore, T.C., 1978. International Cooperation for the Conservation of
Tropical and Sub-Tropical Forest Genetic Resources Exemplified by South East Asia.
Eighth World Forestry Congress. FQL/26-11.
Libby, W.J., Krafton, D. & Fins, L., 1978. California Conifers. In; The Methodology
of Conservation of Forest Genetic Resources. (Ed. L. Roche). FO:MISC/75/8, FAO/UNEP,
Rome.
Lamprey, H.F., 1975. The distribution of protected areas in relation to the needs of biotlc
community conservation in Eastern Africa. IUCN Occasional Paper no. 16. Merges,
Switzerland.
Namkoong, G., 1978. Selecting Strategies for the Future. Third World Consultation on
Forest Tree Breeding. FO-FTB-7 7-6/1. Canberra, Australia.
Namkoong, G. , 1979a. Methods of pollen sampling for gene conservation. Chapter 17.
Pollen Management Handbook. Southern Forest Tree Improvement Committee (en prensa).
Namkoong, G, 1979b. Introduction to quantitative Genetics in Forestry. USDA, Forest
Service Technical Bulletin No 1588. Washington D.C.
Nikles, D.G., Burley, J.& Barnes, R.D. (Eds), 1978. Progress and Problems of Genetic
Improvement of Tropical Forest Trees. Proceedings of a Joint Workshop of IUFRO Working
Parties S2.02.08 and S2.03.01, held in Brisbane, Australia, 4-7 April 1977. Vols. I and
II. Commonwealth Forestry Institute, Oxford, U.K.
Roche, L. (Editor), 1978. The Methodology of Conservation of Forest Genetic Resources.
Report on a Pilot Study. FO:MISC/75/8. FAO/UNEP, Roma.
Sastrapradja, S. ejt al, 1978. The conservation of forest animal and plant genetic
resources. Eighth World Forestry Congress. FQL/26-0.
Sneep, J.& Hendriksen, A.J.T. (Eds), 1979. Plant Breeding Prospectives. Centre for
Agricultural Publishing and Documentation, Wageningen, Netherlands.
Turnbull, J.W., 1978. Exploration and Conservation of Eucalypt Gene Resources. Tercera
Consulta Mundial sobre el Mejoramiento de Arboles Forestales. FO-FTB-7 7-1 /A, Canberra,
Australia.
Wang, B.S.P., 1978. Tree Seed and pollen storage for genetic conservation: possibilities
and limitations. In; Methodology of Conservation of Forest Genetic Resources. (Ed.
L. Roche). FO:MISC/75/8. FAO/UNEP, Rome.
Whltmore, T.C., 1975a. Conservation review of tropical rain forests: General Considera-
tions and Asia. IUCN, Morges, Switzerland.
Whitmore, T.C., 1975b. Tropical Rain Forests of the Far East. Oxford University Press,
U.K.
Wlllan, R.L., 1973 Forestry: Improving the Use of Genetic Resources. Span 16 (3):119-122,
Willan, R.L. & Palmberg, C., 1974. Better Use of Forest Genetic Resources. En: Report
on the FAO/DANIDA Training Course on Forest tree Improvement. Kenya, 1973.
FAO/DEN/TF 112. FAO, Rome.
of Finns Carbaea var . liondmrensl s
C CON ARE, diaguairainas)
- 35 -
ANNEX 1
FOREST TREE GENETIC RESOURCES
Phases and Operations I/
BEGIN HERE
V Willan and
Palmberg
(1974)
- 36 -
4
^
i
^
k
^
^ i
~^
i
A
C
i
I
r
,
i
i
i
*
i
t
i
i
i
t
i
^
i
,
I
^
!
1
2
1
i
'
p. c
I
. j
O
rH
>~
h
CM
j
1
^
,
CM
i
0)
1
J
C
i
i
i
1
V
15
OJ
jq
1
(4
t
i
c2
,._ J
Ll
V
pro
1
*""
^
1
CO fH
rvj rv
S 6
_
i
1
i
QE a
5 M
.
L^ '
^
*
,
1
^j
u
1
^
^"*
S *""!
t
1
^H ^A C
I-
||
,
1
i &
i
r-
i **
!
i
i
I
E **
H
I
2 i
1
a -c
1
S pL
i
'r
1
1
1
i
s
1
i
i
s
1
1
i
1
i
i
1
O Q
1
1
4
i
r
1
j
^
1
"
C
>
S3
6_
^
r
g
s
g
^
4
S
B
S
*H**T
l
i
i
I-
2
i
i
g
i ^
,
sn
s.
I,
h 1
1
8^
4* 1
<
1
B
TEAR
(Tear - beginning
oologioal explorat
*^_ , ^
EXPLORATION
- Botanical
(0-200 years)
t
*
( *1
1
l j
,i
J
I!
f 1
?
3
$
4
3 c
1 H
ii
\
\
q
I
g
i
A
i
E
i
i!
J |
h
vw** na wi*M0n v A A, *w
OFF-SUB COflSEHVATI
- Seed* pollen
i
%
5
I
l
urn HO* Houaanoo
D
51
n
D 1
r
5 <
si
S1
^
1
i
r
i
OBllKrlC UIFROVUNUMr
U Selection of plus
1 4*i i_.t..._1 -4.
5
j-
3.
I
ss
3,
3
i
r-l
a
.a
M
O
- 37 -
Annex 3
PROGRESS IN THE CONSERVATION AND UTILIZATION OF
FOREST TREE GENETIC RESOURCES
Concern over the loss of genetic diversity has grown rapidly since early fifties,
and has spurred increasing national and international action.
Some national institutes had already begun systematic forest tree seed collections
for international use at that time, but the present picture has spurred international and
national exploration, collection, conservation and research efforts in forest tree genetic
resources . These efforts have created world awareness of the need to conserve these
resources, mobilizing national and international funds for experimental plans, pilot studies
and practical activities in this field.
The panel has met four times. FAO has published information on these meetings,
indicating progress, past and present trends, and recommendations for future action (FAO,
1969, 1972, 1975b and 1977). So far, the funds recommended by the panel for programmes
coordinated by the FAO Forestry Department have been mainly earmarked for the exploration
and collection phases to fund certain organizations working in these fields. FAO cooperates
not only with national institutes, but with other international agencies such as Unesco ,
IUCN - , and UNEP $.' , and collaborates actively with the competent IUFRO - working parties.
Recently some funds have been received from the International Board of Plant Genetic
Resources (IBPGR), an auxiliary body of the Consultancy Group on International Agricultural
Research (CGIAR), for the purpose of mobilizing long-term financial support to close the
gaps in agricultural research in the developing countries.
Through the Panel, priorities have been established by region and by species for
each phase of a programme on forest tree genetic resources (FAO, 1977, Appendix 8).
These priorities, which are periodically reviewed in the light of the most recent discove-
ries and measures adopted, are based on the extent to which genetic resources of species
are endangered, and their potential or actual socio-economic importance. However, as it
is only in the course of exploration that exact information will be obtained on the
conservation status of a species, the priorities indicated and species included in the list
reflect to a certain extent the quantity and quality of the information available to the
Panel to adopt its decisions, and not only the real situation (Keiding and Kemp, 1978).
Based on the orders of priority indicated by the Panel so far, full-spectrum ex-
ploration and collection activities have been carried out, followed by the establishment
of centrally coordinated international provenance trials for 12 tropical species. Conside-
rable progress has also been made in the exploration, collection, distribution and evalua-
tion of various genera, including Tectona, Populus, Pinus, Pseudotsuga, Araucaria, and
Eucalyptus. Among the genera most recently included in the Programme are Acacia, Prosopis,
Terminalia and Aucoumea. FAO has published abstracts of the most important collections
(FAO 1975b; FAO 1977) . The reader may also consult the FAO review "Information of Forest
Tree Genetic Resources".
Although many of the species trials based on range-wide collections carried out in
recent years are too new to supply precise data, many already indicate the existence of
major provenance differences and clear interactions between provenance and environment,
confirming that research on provenance is as important for tropical species as it is for
temperate species. The findings of many of these trials have been published in summary
form by species and countries in the proceedings of the joint IUFRO Working Parties S2.02.08
and S2.03.02, held in 1971, 1973 and 1977 (Bur ley and Nikles, 1972, 1973a, 1973b; Nikles,
Bur ley and Barnes, 1978).
I/ United Nations Educational, Scientific and Cultural Organization.
2/ International Union for the Conservation of Nature and Natural Resources.
3/ United Nations Environment Programme.
4/ International Union of Forestry Research Organizations.
8
fi
HI
i
g
3
8
3
S
S
1
i
X 0.
!
o o
8
g
S
OO
8
cu
OS
8
0\
OS
0\
- 39 -
In 1975, with financial support from UNEP, FAO carried out a pilot study which
gave rise to the publication "Methodology of the Conservation of Forest Tree Genetic
Resources" (Roche, 1978a). Based on the technical recommendations of this study and on
the early findings of the exploration and international provenance trials mentioned above,
in 1975/76 an FAO/UNEP project was launched for the conservation of genetic resources of
certain species and provenance. This experimental project, which includes elements of
both on-site and off-site conservation, is now drawing to a close, and a report on its
achievements and conclusions is now being prepared*
The off-site component of the above FAO/UNEP project has performed as expected.
Over the last four years, 33 international ex situ conservation/selection stands, each
10 ha in size, have been established in five African countries and one Asian country, with
a total of 11 provenances of four different species (see Table 1). In addition to the
international stand funded by UNEP and FAO, many of the countries participating in the
project, as well as some neighbouring countries, have established national ex situ selec-
tion/conservation stands. In Appendices 7/1 to 7/4 of the report of the Fourth Session of
FAO Panel of Experts on Forest Genetic Resources (FAO, 1977) include the reasons for the
selection of the species /provenance for the project, the agreement between FAO and the par-
ticipating countries, some recommendations on the establishment, management and handling
of stands, and cost estimates. Succinctly, the agreement calls for international finan-
cing to foot the bill for seed costs, plus the projected cost of the first two years of
the establishment phase. The host government pledges to supervise the establishment, main-
tenance and management of the stands, and to make 50 percent of the seeds, or other repro-
ductive material collected, available to other countries at cost price.
DANIDA (Danish International Development Agency) began a complementary project in
1979 on off-site conservation.
It has proven difficult to reach agreements on in situ conservation. The FAO/UNEP
project only provided funds for two botanical reserves In Zambia for the rn situ 'conserva-
tion of Baikiaea plurijuga (Zambia Sequoia) or Zambesi Redwood. The main reasons why
identifying proper zones for on-site conservation is a problem are listed below:
1) On-site conservation in the tropics generally has to contend with heterogeneous
ecosystems, of which economically valuable species are only a small part. When
funds are scarce, national priorities for expenditures and efforts tend to
centre on other sectors and on other species;
2) Tropical ecosystems are complicated and not enough is known about them, unlike
ex situ conservation stands (even-maturing monocrops). They are also difficult
to manage;
3) It is difficult to predict when the first substantial benefits from in situ
conservation will be produced at the national level;
A) It is often difficult to identify which specific vital operations or phases of
in situ conservation might be suitable for international, short-term financing;
5) The establishment of ex situ conservation stands is a specialized type of re-
forestation, and, as such, is clearly within the sphere of competence of the
forestry service of the country involved. However conservation of the ecosystem
(and therefore in situ conservation) is often under some other authority such
as departments responsible for flora and fauna, national parks, etc.
There has been no let-up in action to overcome these difficulties.
Among the progress achieved in the dissemination of information on forest tree
genetic resources in recent years we may list such meetings as: (1) the three FAO/IUFRO
World Consultations on Forest Tree Improvement (Stockholm, 1963; Washington, 1969; Canberra,
1977), which viewed existing data on the scientific principles of forest tree improvement
and forest genetics; practical advantages and progress in tree improvement; and problems
and prospects for the utilization and conservation of forest tree genetic resources; (11)
the three meetings of the IUFRO working parties S2.02.08 and S2.03.01 already mentioned
in this paper; (ill) the Eighth World Forestry Congress (Indonesia, 1978), which acknowledged
the fundamental importance of genetic conservation, devoting a session to this item on Its
agenda. A series of training courses have also been held on tree improvement.
- 40 -
These were financed by UNDP and DANIDA and organized and implemented by the FAO Forestry
Department (Denmark, 1966; U/S/A/, 1969; Hungary, 1971; Kenya, 1973; Thailand, 1975). A
training course on tree improvement organized by the Forest Research Division of the
Commonwealth Scientific and Industrial Research Organization, Canberra, and financed by
the Australian Government, was held in Australia in 1977. The International Seed Trials
Association, ISTA, has organized various seminars on the processing and analysis of
forest seeds.
The FAO publication "Information on Forest Tree Genetic Resources' 1 (FAO 1973-79)
launched in 1973 publishes three Issues in the biennium. The review features periodic
reports on these and other meetings, seed collections, provenances and the exploration,
evaluation, utilization and conservation of forest tree genetic resources ,
- 41 -
SAMPLING IN SEED COLLECTION
C. Palmberg
Forest Resources Division
Forestry Department
FAO
CONTENTS
Page
Introduction 41
Collection for evaluation 41
Sampling at the population level 42
Sampling at the individual level 43
Collection for conservation 43
Collection for utilization 43
Sampling at the population level 43
Sampling at the individual level ... 44
General Remarks 44
Bibliography 44
INTRODUCTION
In any seed collection, sampling will ultimately determine the proportion of the
genetic variation present that we will capture. Errors or carelessness in sampling* i.e.
in the selection of the populations and the trees from which seed is collected, cannot
be remedied at the planting stage, no matter how careful and sophisticated our experimen-
tal design is or however carefully we establish our plantations.
In broad terms, sampling is done at two levels, the population level (provenance
and stand), and the individual level. The choice of the exact method and intensity of
sampling will depend on the specific objectives of the seed collection.
The main objectives of seed collection are:
(i) Evaluation
(ii) Conservation
(iii) Utilization (large-scale afforestation).
COLLECTION FOR EVALUATION
Evaluation, in this context, generally refers to the establishment of species and
provenance trials in which the range and type of variation is assessed and the adaptation
of the various species/provenances to potential plantation sites is evaluated.
The two main questions to be answered when planning a specific seed collection
mission is how to allocate, in practice, the available time and funds between the frequency
(how many sites?) compared with the intensity (how many trees per site?).
- 42 -
SAMPLING AT THE POPULATION LEVEL
Initially, sampling is done throughout the range of the species (Barner 1974;
Turnbull 1975). If adequate information is not available on the distribution of the
species, searches of literature and herbaria, and contact with foresters, amateur
botanists, and others living in or close to the species range may help define its bound-
aries. Aerial photographs, if available, can sometimes be used to save time in picking
out possible collection sites in unknown country or in difficult terrain (Turnbull 1975).
Ideally, taxonoroic and botanic exploration should precede collection, as efficient
sampling schemes can only be devised based on knowledge of the distribution and ecology of
the species. Sometimes, however, the activities of exploration and preliminary collection
will have to be combined. This kind of a single, combined expedition cannot be expected
to furnish all answers, and a series of reconnaissance and seed collection missions are
necessary.
The number of seed sources (provenances _!/) to be sampled will depend on the
extent and the heterogeneity of the distribution range and on the genetic diversity of
the species. As little will be known about variation patterns of the species in first-
stage studies, sampling can be done on a fairly coarse grid, collecting at rather widely
separated intervals following environmental gradients.
Genetic diversity is generally largest in areas which are optimal for the develop-
ment of the species in question. However, at the limits of the ecological range, outlying
populations of a species may be exposed to extremes of temperature, rainfall or edaphic
conditions. Such provenances may possess morphological and physiological characteristics
which will be of great potential for particular environments. For this reason it is par-
ticularly important that these marginal sites are included in the collections (Turnbull,
1975).
When promising, broad general regions have been located through first-stage pro-
venance trials, second-stage provenance collections should be concentrated on these,
through sampling on a finer network (more collection points in a more limited area).
Examples of sampling schemes used for some specific species are discussed by Turnbull
(1975).
The specific stands from which seed is collected will have to meet certain cri-
teria. A stand has been defined as a sufficiently large population of trees possessing
sufficient uniformity in composition, constitution and arrangement to be distinguished
from adjacent crops (OECD 1971). In practice, the main consideration should be 'to select
a population which is large enough to allow for sufficient cross-pollination between a
large number of trees and which is isolated from related species to minimize risks of hy-
bridization (Turnbull 1975; Melchior and Venegas Tovar 1978). The application of such
criteria to some tropical species which are found as isolated trees in mixed forest may be
difficult. In such instances, individual tree lots may be kept separate or combined with
others covering a considerable area to represent f a stand' (Turnbull 1975).
Melchior and Venegas Tovar (1978), referring to Eucalyptus globulus plantations
in Colombia, consider 300 individuals a minimum number for a stand suitable for seed
collection purposes.
Stands should be of such an age that seed is produced in quantity, due to prac-
tical as well as genetic considerations.
In the case of introduced species, seed should whenever possible be collected from
stands of known origin 2/. The general history of the stand is relevant in the case of
both indigenous and introduced species. Any treatment which may have altered the distri-
bution of the phenotypes should be noted and stands which have been selectively thinned
to remove the best phenotypes should be avoided (Turnbull 1975).
I/ Provenance: The place in which any stand of trees is growing. The stand may be
indigenous or non -indigenous (OECD 1971).
II Origin: For an indigenous stand the origin is the place in which the trees are
growing; for a non-indigenous stand the origin is the place from which
the seed or plants were originally introduced (OECD) 1971.
- 43 -
SAMPLING AT THE INDIVIDUAL LEVEL
In selecting trees for sampling for provenance studies the aim is to take a sample
which is as representative as possible of the population. The main considerations in
sampling are the number of trees, the type of tree and the distance between trees to be
sampled. In order to give maximum flexibility in selection it is desirable that seed
collections take place when there is a good seed crop on the majority of trees (Turnbull)
1975)* Good seed years will also furnish a seed sample which more fully represents the
population genetically.
Sampling at the individual level can be done either for variability or for superio-
rity in specific characteristics. Provenance studies are designed to expose genetic
differences between populations and to indicate the best localities for seed collections.
Such research can be done as well, if not better, with seed from random trees as with seed
from carefully selected phenotypically superior individuals. IUFRO standards suggest
selecting 'average or not less than average 1 trees of dominant or co-dominant status in
'normal 1 as compared with 'plus 1 stands.
In sampling trees for provenance trials from natural stands, the question of
spacing between sample trees is important because of the need to avoid trees which are
closely related genetically(half-sibs) or trees which have an abnormally high incidence
of self-pollination. Although recommendations vary, it is generally considered that seed
trees should be between 100 and 300 m from each other to avoid narrowing down the varia-
tion sampled due to relatedness or inbreeding (Turnbull 1975; FAO 1975; Melchior and Venegas
Tovar 1978). Adjacent trees in plantations are usually not closely related because the
seed has been bulked and so there need be no restrictions on sampling from adjacent trees
(Turnbull 1975).
The number of trees sampled per stand will vary according to the species and its
breeding system. Ten to twenty-five trees can be considered a minimum in species" which
are found in stands (Barnar 1974); possible solutions in dealing with tropical species
which occur as individuals rather than as stands have been discussed above.
The number of seed required per tree for provenance trials need not be large:
10 000 seeds per tree would be sufficient for each experiment, provided seeds are picked
out one per container or individually sown in seedbeds.
COLLECTION FOR CONSERVATION
The main consideration when sampling for conservation ex situ is the maintenance
of maximum allelic diversity. Strategies are discussed in the lecture 'Principles and
Sstrategies for the Improved Use of Forest Genetic Resources'.
COLLECTION FOR UTILIZATION
SAMPLING AT THE POPULATION LEVEL
Collection for utilization is generally done from a limited part of the range,
from populations which have been found through provenance trials to be well-adapted to
the environmental conditions of the planting site. However, as many countries will start
their planting programmes before sufficient information is available on provenance
performance, climatic and edaphic matching is often done toselect the most likely seed
sources. As most species and provenances have some degree of plasticity (i.e. capacity
to adapt to environmental conditions which differ somewhat from the natural ones)
(Willan 1979), this can be an acceptable temporary measure provided that full-scale
evaluation trials are started in parallel. Systematic evaluation is always necessary,
as other factors than climate and soil in the present environment will determine the
genetic make-up of trees, making the performance of new introductions in a matching
environment unpredictable.
If collection is carried out in plantations or in even-aged, homogeneous stands,
the genetic quality of the seed can generally be somewhat improved by sampling for superi-
ority of stands as well as of individual trees (Turnbull 1975). However, it should be
noted that in the case of uneven-aged natural stands and stands in which past history is
- 44 -
not known, phcnotypic performance is no guarantee of genetic superiority: the probability
of superiority being due to purely environmental causes is, in fact, 50% (Melchior 1977;
Barner 1974).
SAMPLING AT THE INDIVIDUAL LEVEL
The lowest acceptable standard of collection at the individual level 9 often used, by
necessity f to meet the needs for substantial quantities of seed for large-scale afforesta-
tion programmes, involves the collection of seed from all trees of a particular provenance
or provenances t except from those of notably inferior phenotype (Turnbull 1977). Ideally,
the trees sampled within a stand should be dominants, free of pests and diseases and in the
case of plantations, of better than average form.
In stands regenerated by natural seed fall, collecting from trees growing closely
together may result in plantations with a narrow genetic base, thus responding in a uniform
manner to environmental pressures such as diseases or unforeseen adverse conditions, and
generally having less flexibility to adapt to the requirements of a new site. In addition,
if seed stands are subsequently established in these stands or seed is collected from them,
severe inbreeding effects may occur in the resultant second generation stands. In natural
stands, the distance between trees collected should, as in collections for evaluation,
ideally be 100-300 m and as a minimum requirement, trees should be selected at an interval
that exceeds normal seed fall (Turnbull 1975).
When collecting bulk quantities of seed there is no upper limit for the amount of
seed collected per tree, provided that the number of trees collected is large, When
collecting from standing trees, the lower limit of seed per tree would be set by economic
considerations.
GENERAL REMARKS
Practical limitations and economic possibilities often modify the ideal sampling
strategies outlined above. Factors such as accessibility of stands and yearly fluctua-
tions in the seed crop at the provenance as well as at the individual level also influence
decisions on sampling. If, however, the genetic and biological principles on which
recommended sampling methodologies and intensity are based are known, these modifications
can often be accommodated without too wide-ranging adverse consequences. Detailed record-
ing of sampling procedures used and number of provenances, stands and individual trees
collected from, as well as selection criteria in each instance, are of fundamental impor-
tance and cannot be over stressed.
BIBLIOGRAPHY
Barner, H., 1974. Classification of Sources for Procurement of Forest Reproductive
Material. In; Report on the FAO/DANIDA Training Course on Forest Tree Improvement,
held in Limuru, Kenya, September-October 1973. FAO/DEN-TF-112. FAO, Rome
FAO, 1975. Forest Genetic Resources Information No. 4. Forestry Occasional Paper 1975/1.
FAO, Rome.
Melchior, G.H., 1977. Program* Preliminar de un Ensayo de Procedencia de Cordia alliodora,
Cupressus lusitanica y otras Especies Nativas y Exoticas. Proyecto Invest igaciones y
Desarrollo Industrial Forestales COL/74/005. P1F No. 7. Bogoti, Colombia.
Melchior, G.H. & Venegas Tovar, L., 1978. Propuesta para Asegurar el Suministro de
Semi 11 as de Eucalyptus globulus en Calidad Comercial y Geneticamente Mejoradas. Projecto
Investigaciones y Desarrollo Industrial Forestales COL/74/005. PIF No. 14. Bogoti,
Colombia.
OECD, 1971. OECD Scheme for the Control of Reproductive Material Moving in International
Trade. Paris, France.
- 45 -
Turnbull, J., 1975. Seed Collection - Sampling Considerations and Collection Techniques.
In: Report of the FAO/DANIDA Training Course on Forest Seed Collection and Handling,
held in Chiang Mai, Thailand, February-March 1975. FAO/TF/RAS-11 (DEN). FAO, Rome.
Turnbull, J., 1977. Seed Collection and Certification. In: Selected Reference Papers,
International Training Course in Forest Tree Breeding. Australian Development Assistance
Agency. Canberra, Australia.
Will an, R.L., 1979. Eucalypts for South-East Asia. Tropical Agriculture Research Series
No. 12. Ministry of Agriculture, Forestry and Fisheries, Japan.
- 46 -
COLLECTION AND HANDLING OF FOREST SEEDS
C. Palmberg i / and G.H. Melchior - /
CONTENTS
Page
Introduct ion 47
Biological pre-conditions for flowering and fruiting 47
The juvenile-adult phase. 47
Actual population size 48
Planning seed collection 49
Determining seed requirements 49
Harvest forecast and evaluation % 49
Sampling in seed collection 50
Seed collection techniques 50
On-ground collection 50
Collection from standing trees 51
Training and safety 53
Packing and labelling collected seeds 53
Drying, post-harvest ripening and seed cleaning 53
Supervision 55
Bibliography 55
Annex 1. Equipment needed for seed collection, information on area and sample
for herbarium 57
Annex 2 and 3. Sample seed collection data sheet 58-59
y Forest Resources Division, Forestry Department, FAO,
Via delle Terme di Caracal la, 1, 00100 Rome, Italy
2/ Bundesforschungsanstalt fur Holz und Forstwirtschaf t ,
Institut fur Forstgenetik und Forstpf lanzenziichtung
Sieker Land Strasse 2, D-207 Grosshahnsdorf 2
The Federal Republic of Germany
INTRODUCTION
A foremost constraint of large-scale planting of native or exotic species in Latin
America, as in other parts of the world, is the lack of available information on basic
theory and practical methods of harvesting, handling and treating forest seeds. Lack of
contacts with the outside and especially lack of foreign exchange mean that the seed gap
often cannot be closed by purchases abroad*
Since the Latin American countries still have relatively extensive forested areas,
domestic requirements for seeds of native species could be met by collections in local
stands. There is an urgent need to establish seed areas or plantations to produce exotic
seeds so that natural self-reliance in seeds can be achieved as soon as possible.
Forest seeds may be collected to:
- ensure the continuous short and long-term supply of reproductive material for
plantation programmes. Seed collection to be determined by the demand for final
products within the country;
- guarantee the reproductive material needed for scientific trials;
- supply the reproduction material for the establishment of ex situ gene banks,
and for the establishment and expansion of a r bo re turns, botanical gardens and
other collections of tree species;
- supply reproductive material for beautif ication of landscapes, recreational
sites, cities, highways and the like.
Reproductive material collection methods vary in accordance with the above-
mentioned objectives and will depend on the species in question, the sites to be re-
afforested, the type of reproductive material to use, and so forth.
BIOLOGICAL PRE-CONDITIONS FOR FLOWERING AND FRUITING
Detailed discussions of the formation of gametes and zygotes in conifer and hardwood
tree species can be found in the Mittak manual (1978) and in various biology texts. This
lecture will therefore be confined to pointing out certain biological aspects of particular
importance for the development of forest seeds.
THE JUVENILE - ADULT PHASE
The zygote, seed and sporophyte are formed through the union of masculine and
feminine gametes. A diploid individual is produced by mitotic cell division. This
individual can be defined in terms of its specific morphological and physiological features
such as: the position of the needles, shape of the leaves and duration of the juvenile
period.
The characteristic feature of the juvenile period is that the individual is incapable
of sexual reproduction. The length oft the period is further influenced, in addition to
hereditary factors, by environmental circumstances. The duration of the period can range
from 2 to over 50 years, depending on the species. Cordia alliodora in lowland areas for
instance, matures in 4-6 years, Podocarpus in mountain areas requires 30 years or more, and
Quercus in the temperate zone over 50 years.
When the tree reaches maturity its morphological characters change. The initiation
of flowering and fruit setting is one criterion closely related to the degree of maturity
of the tree. Unfortunately we know very little about phenological variation among and
within tropical forest species and about the genetic and environmental influence* But
it is known that individual trees within a specific stand or forest usually vary as to age
at first flowering and the regularity of flowering and fruit setting, especially when young.
One result of this variation, aggravated by the typically great distances between individual
trees or groups of individuals of the same species in tropical forests, is the birth of
subpopulations which all flower at the same time, if these subpopulations are small there
- 48 -
is a grave danger of self ing among them with all the adverse effects of inbreeding for the
next generation: instability, low resistance to biotic and abiotic factors, and reduced
vigour. This is why it is absolutely necessary to resist the temptation to collect seeds
in very young stands, despite the fact that collection would often be easier and less
expensive in such stands than in mature ones* Only when 60 to 100 percent of the indivi-
duals in a given population have flowered (only adult populations will achieve such per-
centages) should seed collection for commercial plantations be undertaken.
To speed up the onset of the adult phase of the tree and thus speed up flowering
and fruit setting, stands for seed collection are usually established in sites with fertile
soils and optimum climates for the flowering of the specific species. Sometimes it is also
possible to speed up the adult phase and increase flowering and fruit setting by applications
of fertilizers or plant hormones.
None of the above-mentioned methods should be used unless accompanied by careful
observations or prior experimentation to throw light on the specific characteristics and
requirements of the species.
ACTUAL POPULATION SIZE
By population size is meant the proportion of genotypes actually participating in
fertilization and seed production within any given year.
Ideally, to enhance the physiological and genetic properties of seeds, all gametes
in populations with a large number of unrelated individuals should combine at random.
This is because crosses with foreign pollen are a positive factor in selection, as they
maintain allele heterozygosity and suppress the appearance of sub-lethal alleles.
The following factors affect actual population size:
(i) absolute population size;
(ii) mechanical barriers, which might consist of: natural windbreaks (reducing the
movement of wind-borne pollen) or a very large mixture of non-hybridizing
species;
(iii) phenology and synchronization of flowering. Non-synchronized flowering can
combine with scant flowering some years to drastically reduce actual population
size and hinder continuous seed supply and progress in breeding programmes.
Synchronization will affect the genetic make-up of the seeds collected, as
genetic composition usually varies within the stand from one year to the next
depending on which trees are flowering. It will often be necessary in
breeding programmes to establish a series of seed orchards, each having a pheno-
logically synchronized subpopulation;
(iv) intra-genotype incompatibility. This phenomenon, which is usually a barrier to
selfing, in particular, and fertilization among closely related trees can be
based on the following factors:
- the pollen does not germinate on the stigma;
- the pollen tube does not penetrate the stigma;
- the pollen tube interacts with the tissues of the style and does not reach
the ovule;
- the gametes do not function due to incompatibility;
- gamete union takes place but the zygote dies as the endosperm is incompatible
or else, in polyembryonic species, selection operates against the gamete.
- 49 -
PLANNING SEED COLLECTING
DETERMINING SEED REQUIREMENTS If
Planning for reforestation must include seed procurement and ensure a sufficient,
constant supply of seeds,
The person in charge will calculate seed needs for forest nurseries or direct plant-
ing in terms of the size of the projected plantation, the geographical area and the species
one plans to use for reforestation. If the plan is to reforest 1 000 ha with Pinus oocarpa,
at spacings of 2.5 x 2*0 m, for example, the number of plants required will be 2 000
plants/ha. Therefore 2 000 000 plants will be needed to reforest 1 000 ha. Two
good quality seeds are needed to produce one plant in a nursery bag in the forest nursery.
The number of viable seeds needed for the planned reforestation will therefore be:
2 x 2 000 000 = 4 000 000
The number of viable seeds in a specific volume of cones varies greatly with the
species, and even within the same species. Assuming an average 48 000 Pinus oocarpa seeds
per hectolitre of cones, the number of seeds in the example quoted above is
4 000 000 : 48 000 = 83.3 hi.
It therefore follows that approximately 83 hi of mature P. oocarpa cones must be
collected to reforest 1 000 ha (plants in nursery bags).
Nonetheless, as it is unlikely the required number of seeds can be collected every
year, the largest possible number must be collected in good years and kept in cold storage
for the future. The provision for "x" number of years to be collected in a planned harvest
will depend on such factors as flowering period and the storage life of the seed under
existing storage conditions. It is a good idea to store enough seeds for at least two
years plus the present year. In this instance, the amount of seed to be collected would be:
3 x 83 = 249 hi
Yearly records of production cone maturity dates for each species, and provenance
are important, good planning and data on the periodicity of good harvest are dependent on
such records.
I/
HARVEST FORECAST AND EVALUATION^
When the reproductive buds have formed, the collection planner or supervisor can
use the appropriate sampling method to determine whether a potential harvest is develop-
ing. We stress the word potential because the many impediments to seed development can
destroy the reproductive structures or maturing seeds at any stage of the process.
Early forecasts are based on sampling. Female buds are collected as samples from
10-20 trees. Buds for identification are collected three branches from the top of the crown.
A statistical evaluation of the sample material will give an indication of harvest potential.
Theoretically, the greater the production of female buds, the greater the probability of a
good seed harvest. Unfortunately, the quantitative relationship between female bud counts
and potential harvest sice is not completely reliable. But comparisons between areas and
periodicity 4re always of some practical value* The procedure described may be useful until
such time as a more suitable technique is devised for forecasting harvests of a species
important for reforestation.
The forecast of subsequent forecasts may be based on a female flower count or, later,
a count of new cones.
I/ This part of the lecture is based on Mittak (1978).
- 50 -
The most common error in classifying cone harvest! is to count old cones which have
scattered their seeds in earlier seasons, and to use trees growing along roads as a basis
for estimates. Such trees receive more sunshine and frequently have more cones and buds
than those within closed stands.
Based on the qualifications of the various possible seed collection areas, the planner
will assign collection priorities subject to the next seed yield evaluation (content) of
the harvest.
The main objective of the harvest evaluation is to indicate the number of healthy
seeds in cones. Cones are picked at random from different trees throughout the stand. Care
must be taken to ensure that the samples are representative of all cones, i.e., they must
be collected evenly all over the crown of each seed tree.
Some ten representative trees in the stand are selected for evaluation and ten cones
collected from each. The ensuing one hundred cones are split lengthwise down the middle for
examination. All the good seeds on one of the cut surfaces are counted and then minutely
examined. They are then treated separately in a 65C oven until the pine cone sections
open releasing the seeds. The total number of good seeds from the cone is divided by the
number of seeds counted on a single face of the split cone. An average figure can be
worked out for the one hundred cones once these figures are available.
Seed pests and diseases not only affect cone or fruit ability to produce mature seeds,
they also affect subsequent seedling development. The usual estimate is that if more than
50 percent of the seed is damaged, seeds should not be collected from that stand.
Together with the quality and number of seeds it is important to get some indication
of how ripe they are so a date can be set for harvesting.
The seed incision test, used in conjunction with other indicators of maturity, is
usually very helpful in determining maturity. The test consists of cutting each seed
lenthwise exactly down the middle with a razor blade. The contents of 20-30 seeds are
then examined under a 10 power lens. The normal process is that as the seeds mature, the
embryos elongate and turn yellowish whilst the endosperm turns from milky and sticky to
firm consistency (similar to coconut meat). The tegument and wing also darken. Usually, the
embryo has to have elongated at least 75 percent of its potential length to ensure seed
viability. The possible or potential length of the embryo is the length of the cavity
within the endosperm.
The incision test should not be performed until 3-4 weeks before seed maturity as
before then the unfertilized ovules of some species can simulate normal development. In
other words, they contain endosperm and look like good seed. However, since they have not
been fertilized, there is no embryo and they cannot form viable seeds. Some three weeks
before maturing, most of the unfertilized ovules stop developing and their endosperm tissue
will dry up and form empty pods.
SAMPLING IN SEED COLLECTION
See previous lecture.
SEED COLLECTION TECHNIQUES
ON-GROUND COLLECTION
Seed collection from natural seed fall, being a low-cost, easy method, is very
widespread. It includes collecting seeds of fruits which have fallen from standing trees
and collecting from felled or fallen trees*
This method is frequently used for fairly large seeds or fruits, including, in
temperate regions Quercus spp., Fagus spp., Castanea spp.; in tropical areas Tectona spp.,
Sb&Cfii spp* Triolociiiton spp. and Gmelina arborea (Turnbull 1975; USDA 1974). Where
possible it is recommended that the ground under the trees be cleared prior to collection.
- 51 -
Although seed collection from natural seed fall is a relatively easy, economical
method, it does have certain serious drawbacks. The seed viability of many species such
as Shorea spp., is very quickly lost once the seeds have been shed. Seeds on the ground
are also very prone to damage by insects, fungi and animals. It is therefore imperative
that shed seed be picked as soon as possible, remembering that the first seeds or fruits
to fall are usually of rather poor quality (Turnbull 1975). Another drawback with this
method is that it is normally not possible to say from which tree the seed was shed and
there is, therefore, no indication of phenotype quality.
A closely related method which does eliminate many of the above drawbacks is to use
tree shakers. These consist of a hydraulic arm mounted on a tractor which is used during
the collection season to shake the trunks to make the ripe fruit fall (Turnbull 1975;
Ottone 1978).
Animals such as squirrels sometimes gather cones or seeds. Foresters sometimes raid
these stores for quick 'seed collection* It is a very common method e.g. for Pseudotsuga
menziesii in the United States (Turnbull 1975).
Collecting seeds from felled or fallen trees in thinnings or clear-cut areas or
from trees grown specifically for seed collection is another common harvesting method.
Though in theory it is possible to limit the collection to desirable stands using this
method, one can never again collect seed trom tne same tree or stand. Cutting down trees
just to collect seeds in shortsighted and wasteful.
COLLECTING FROM STANDING TREES
Seed collecting from standing trees is the most common large-scale method for
collecting forest seeds. It is a sure method, but careful attention must be paid to follow
the proper safety measures, use the right equipment, and' keep the equipment in good
condition.
They are two distinct harvest methods (i) collecting from standing trees without
climbing, (ii) collecting from standing trees by climbing them.
(i) Collection from standing trees without climbing
There are various ways of harvesting seeds without climbing. Some use equip-
ment to get the fruits and seeds down, such as long, light poles for striking
and shaking branches, knives with telescopic or long handles for cutting fruit,
collection sticks, pole proners and the like.
In Australia a rifle is used to sever the fruit-bearing branches of trees.
Another possibility is a flexible saw operated by two collectors from the base
of the tree. This last method can be used to cut branches up to 20 cm in
diameter (Turnbull 1975).
(ii) Climbing standing trees for seed collection
Tree climbing equipment varies with the species and environmental surroundings.
Burley and Wood (1978) lists of equipment needed for seed collection are found
in Annex 1.
Climbing irons
The usual way to climb trees is to use climbing irons. The forged iron is fastened
to the foot gear with two sturdy leather straps* The foot gear must be solid and firmly
attached to the climber's foot and leg. The iron ends in a fixed spur of varying length
depending on the tree climbing method used. One of the best has a short hook which does not
extend past the sole of the boot. This enables the climber to walk on the ground without
difficulty. Irons can damage trees with thin smooth bark and should therefore not be used
for climbing young or thin barked trees (Mittak 1978; Turnbull 1975). To facilitate the
climber's task, ropes can also be snaked around the trunk of the tree (Mittak 1978).
- 52 _
The IUFRO team* collecting North American conifer seeds used as standard equipment
climbing irons, safety belt, a hook to bend back branches, safety hat and overalls
(Turnbull 1975).
Tree bicycle"Baumvelo"
The use of a Baumvelo or tree bicycle is recommended for collecting seeds, cones or
fruits from trees with very thin, smooth bark. Tree bicycles are very easy to use on trees
with smooth boles. The tree bicycle is particularly used to collect buds and pollen from
plus trees to avoid damaging them (Nittak 1978).
The bikes consist of two pedals supported by a vertical arm. The pedal is fastened
by special straps to the operator's boots. The arm is connected to a steel band forming a
circle of adjustable diameter around the stem of the tree to be climbed. The pedals must be
about 5-8 cm greater in diameter than the tree for ease of movement. In climbing, the
operator puts all his weight on one boot-pedal, then raises the other foot with the other
boot-pedal as high as possible. He then moves the other foot up to meet it. Between the
two operations, he roust adjust his safety belt as high as possible. When he reaches the
crown he can take off the pedals and fasten them to the trunk for freedom of movement. He
must again strap on the pedals to climb down the tree. The tree bicycle is easy to
transport. Climbing with the tree bicycle is easy and rapid after some practice (Mittak
1978; Turnbull 1975; Ottone 1978).
Portable Ladders
Ladders for collecting forest seeds vary greatly in construction and the materials
used. Ladders are a quick, safe method to climb trees up to 15 or 20 ra or even higher.
Mittak (1978) lists the following kinds of ladders for collecting forest seeds:
Rope Ladders, such as sailors use. A rope is shot over a sturdy branch to haul up
the rope ladder which is then made fast to the branch.
One-legged ladders, with alternate short rungs on both sides of the supporting pole.
These ladders are appropriate for uneven terrain and branchy trees. One-legged ladders are
attached by chains to the trunk of the tree. There is usually a bolster between the tree
and the pole ladder at chain height to facilitate access. They are easier to steady than
regular ladders.
Scaling ladders in several sections, usually of aluminium, are needed for climbing
tall trees. Sections vary in length (2-4 m), but for easy carrying no section should
weigh more than 3-4 kg. The sections are designed to fit into one another and are attached
by a chain to the trunk of the tree, which the operator grasps as he climbs.
In addition to the above-mentioned ladder types, ladders mounted on tractors are
used, as well as hydraulic platforms of various types. Often these platforms are only
used in seed orchards since per unit collecting costs are very high (Turnbull 1975).
Nets and other various methods
A special device for climbing trees, especially small trees such as cypress, is the
net built by the United Kingdom Forestry Commission. A triangular net is suspended by
special ropes and snatch blocks from the crown of the tree. The operator can climb up the
net to collect the small fruits which cannot otherwise be harvested without partially
cutting branches, doing great damage to future harvests. This device produces good results
in species like cypress, i.e. species with many small cones or fruits on the outer part of
the crown (Mittak 1978).
Mention has also been made of the use of pulleys attached to a sturdy fork. Even
balloons have been used in different parts of the world, but they are hard to use and the
end result is generally disappointing (Ottone 1978; Turnbull 1975).
- 53 -
TRAINING AND SAFETY I/
IB is very important to train climbers in the various phases of seed collection and
the correct use of tools and equipment.
Foremen should know which trees will produce good seed and which will not, and mark
them prior to harvest.
Personnel safety is, after genetic and biological know-now, the most important aspect
of seed collection.
Personnel should receive practical and theoretical training in accident prevention
and first aid.
Each team should be equipped with a first aid kit and stretcher. Tools, safety
belts, safe hats, protective goggles, ropes, hooks, spurs, ladders, the first aid kit and
all utensils for seed collection should be minutely examined before the climb or before
beginning work. Any tool or other equipment which gets damaged on the job should immediately
be reported and a request made for repair or replacement.
Equipment should be clean. A product should always be brought along to dissolve
resin and other sticky substances like pitch once the Job is finished. Equipment
maintenance promotes safety and makes the equipment last longer.
PACKAGING AND LABELLING COLLECTED SEEDS
After picking, dry cones should be packed in coarse sacks or in solid-bottomed fruit
crates. It is not advisable to use plastic bags for this type of cones. Air circulation is
inadequate and moisture condensation can affect the fruit and seed. The temperature also
rises within plastic bags, which can cause various kinds of damage such as lowered germina-
tion rates, mildewed seeds and fewer seeds released.
The coarse sacks are not completely filled. They are tied with string but not sewed,
so the work will go faster. The reason the sacks are not completely filled is to avoid
tightly packing the cones while they are still in the field and during the pre-drying
period to prevent mildew from moisture and lack of air circulation. The sacks should be
protected from rain and rodents and covered with canvas tarps. The sacks should be moved
and turned two or three times a day and should not touch* The ground or floor should be
dry. If crates are used, they can be piled up to a height of 2-3 m. The corners at the top
of the crate and slatted sides promote air circulation. This stage of the process
constitutes a sort of pre-drying phase affecting subsequent work and final seed quality*
The cones can also be more easily moved to the processing site.
Sacks and crates are also a way to keep the cones during the collection/drying
period* The harvest period is short and so a large number of cones is being handled. Thus
packed, they will keep perfectly (Ottone 1978).
Some species of fruits lose viability if not kept moist. They must be collected
before dry and transported to the temporary storehouse as soon as possible. Plastic bags
can be used for some species for very short times. Sturdy paper bags with or without an
inner plastic lining are also used (USDA 1974).
To guard against lost identity of seed lots it is extremely important to identify
and label them. Each sack or package should be labelled inside and out. The labels must
be water- and moisture-resistant and will list data on the species, provenance, collection
date, name and last name of the collector and so forth. A seed lot number identifies them
on a more detailed data sheet (see Annexes 2 and 3 for sample sheets).
DRYING, POST-HARVEST RIPENING AND CLEANING SEEDS 21
The cones or fruits should be transported as quickly as possible from the temporary
storage area to the final handling and storage centre*
17 ThU part of the lecture is based on Mittak (1978)
|/ This part of the lecture is based on Ottone (1978)
- 54 -
The first step after transport is to dry the cones and extract the seed. There are
two mains drying methods: sun and air drying, and artificial drying.
The natural drying process involves less risk of lowering seed quality but takes
longer. Nor can the process be used in very wet or rainy climates. For easier handling
of cones and seeds and quicker cone drying, the cones and seeds can be piled onto canvases.
The canvas is folded at night to protect the cones and seeds from dew or rain. Rough sheds
in which the air circulates freely are another answer. Another possibility is trays which
are shaken to make the cones release the seeds. Plastic roofs can accelerate the sun
drying process. Tin roofs are not used as they cause moisture condensation.
Increased air humidity in the drying area can produce a condition known as case
hardening ( the same can occur in oven drying). What happens is that the cone reabsorbs
water and the opening process is inhibited to the point that the cone closes or remains
partially closed, preventing seed release.
Humidity and temperature are easier to control with artificial drying. Artificial
drying is quicker (from 6-15 hours) but it does require equipment and installations which
can be very costly. It should be considered that the equipment is only used a few days of
the year. Artificial drying is only recommended for cold, wet areas. Drying should be
done in the shortest possible time to avoid damage to the seeds. The air must be hot and
dry for even, rapid drying. Case hardening can occur if the cone is reheated. Cones must
be pre-dried. Good oven temperature control is essential. The cones should be dried at
the minimum temperature for as short a time as possible. Ideal temperature range is 10C
at the beginning of the process and no higher than 52 W C at the end.
High temperatures will cause physical and physiological damage to the seed as will
humidity (which should be eliminated as quickly as possible). The p re-dry ing operation aids
and shortens oven drying. Pre-drying is therefore always recommended - for how long
depends on the tree species. If rotary ovens are not used, once the fruits have dried they
can be put into a 60 x 150 cm wire drum ( approximate measurements), which is rotated to
detach the seeds not released during the drying process. The same should be done with sun-
dried cones. Small quantities of seeds can be threshed by hand into a crate, or onto a
tarpaulin or brick or wooden floor.
Eucalypt capsules release their seeds in a few hours or at most one day after
picking. The best way to collect eucalypt seeds is therefore to spread them on canvas
tarpaulins sheltered from rain. The seeds can easily be separated from the fruit, branches
and leaves by straining them through a fine mesh screen. They are then packed in closely*
woven cloth sacks.
Fleshy fruits usually have to be soaked to remove the external fleshy covering. The
seed is then removed by hand or with special equipment. The treatment includes alternate
phases of washing, drying and cleaning and must be begun immediately so the fruits will
not have time to ferment (USDA 1974).
One limiting factor in collecting the fruits of various forest species is the time
period in which harvesting is possible. The cones or fruits of some species can be picked
before ripe to lengthen the collection period. They must be stored in such a way that they
can complete their anatomical and biochemical development. The usual post-harvest ripening
temperature, depending on the species, should be 5-1 8C. Constant vigilance is necessary
throughout the process to ensure that the above conditions are met and no problems arise
which might adversely affect the fruit and seeds, such as mildew.
It is also important to know how early the fruits are to be picked. Certain pine
species can be picked when they have reached 1.1-1.2 of their specific weight, or approx-
imately 15-20 days before the normal collection date.
The fruits or cones are stored as above for one month during which time they complete
the development process. They are then dried in the usual way. The wings of winged seeds
must be removed after drying.
To remove the wings the seeds are placed in a wire drum containing two or three sets
of rotary brushes, regulated in accordance with the size of the seed to be processed. These
brushes are attached to a central axis driven by an electric motor. As the drum turns,
- 55 -
the seeds and wings are pressed against the wire and the wings are rubbed off.
The brushes must also be set to leave enough space between the brush and the mesh
to avoid excessive rubbing which would raise the temperature, harming the seed.
If winged seeds are to be sown immediately after, they can be slightly moistened.
The wings will swell with the absorbed moisture and can easily be removed. If seed treated
in this way is to be stored, it must be dried before storage.
Ordinary wire sieves with wooden frames can also be used to remove the wings from
the seeds. These are vegetable seed cleaners with mesh sizes calibrated to seed size.
Seeds can also be winnowed directly out-of-doors or in variable draft winnowing fans.
Whatever method is used for winnowing, the operation also eliminates sterile seeds.
Seeds should not be over-winnowed to avoid eliminating good seeds which happen to be lighter
because they are smaller. It would be better to keep some sterile seeds than to lose good
seed.
At all stages of handling cones, fruits and seeds, it must be remembered that they
are living organisms. Rough handling will lower germination rates and adversely affect
seedling development*
SUPERVISION II
Before leaving the field, the supervisor and foremen must check all the equipment.
In addition to checking the equipment, the supervisor is also in charge of transport, fuel
and daily subsistance allowance for his team. A successful harvest, staff safety and the
cost of seed collection are largely dependent on how well the operation is organized.
Every day the supervisor marks the trees in the area from which seed is to be
collected. The criteria is seed maturity.
It is recommended that harvesting be done by work quotas, first determining costs in
accordance with the amount to be harvested and reception facilities. Inspectors should en-
sure strict compliance with quality standards.
The supervisor must also keep a written record of the output of individual collec-
tors, listing the number of cones, seeds or fruits collected, the operation, the species,
provenance and site number and any other necessary observations. He must also fill out the
collection sheets (see Appendixes 2 and 3).
It is recommended that a random check of one sack in five from each collector be made
to ensure that the fruit or seed collected has been properly identified, that the sack is
not over-full and that good-quality seed has been picked.
Supervision is essential at all subsequent phases of handling to avoid damaging the
seed and ensuring that each seedlot is properly identified. Any treatments must be listed
on the corresponding sheets for each lot.
BIBLIOGRAPHY
Burley J. & Wood, P.J., (1979). (Editors) Handbook on Species and Provenance Research
with Special Reference to Tropical Areas, Commonwealth Forestry Institute. Tropical
Forestry Papers No. 10 & 10A. Oxford, U.K.
Mittak, W.L., (1978). Manual 2. La recolecci6n de semi lias forestales. INAFOR BANSEFOR
FAO/TCP: Proyecto GUA 6/01-T Institute Nacional Forestal, Guatemala Ciudad.
Ottone, J.R., (1978) Recoleccion y tratamiento de frutos para obtener semi 1 Us forestales.
Institute Forestal Nacional. Folleto Tlcnico Forestal No. 45. Buenos Aires.
T7 This part of the lecture is based on Mittak (1978)
- 56 -
Turnbull, J.W., (1975). Seed Collection: Sampling considerations and collection tech-
niques. In: Report on the FAO/DANIDA Training Course on Forest Seed Collection and
Handling, held in Chiang Mai, Thailand, February - March 1975. FOR . TF-RAS / 1 1 ( DEN ) .
US DA, (1974). Seeds of Voody Plants in the United States, Chapter V. U.S. Department of
Agriculture, Forest Service. Agriculture Handbook 450. Washington D.C.
****************************
- 57 -
Annex 1
EQUIPMENT NECESSARY FOR SEED COLLECTION, SITE
INFORMATION AND SAMPLES FOR HERBARIUM -/
A - Seed collection
Containers for seeds (in field). Sacks and bags (may "be re-usable).
Containers for seeds (for shipping). Cotton and canvas bags (sent with seeds).
Tags for trees, e.g., plastic rings.
Climbing equipment. Climbing irons, tree bicycles or ladders. Safety belt,
safety harness, safety helmets, tool -carrying harnesses.
Seed cutters, e.g. hooks and rakes for cones, pruning poles, hand shears.
(Heavy) plastic tarps to protect fruits, extract seeds and so forth.
Binoculars to examine tree crown, check fruit development, etc.
Portable walkie-t-t alkie (special permit may be necessary).
Insecticide and fungicide dusts for seed protection (use with caution).
Axes, saws, machetes, knives.
Harnesses, ropes, labelling equipment and labels.
B - Site description
Notebook, site description sheets.
Maps (including copies with major outlines to be filled out in detail).
Compass.
Altimeter.
Meteorological equipment (hygrometer, maximum/minimum thermometer).
Soil survey equipment (core drills, soil colour charts, pH test).
Tree measuring equipment (altimeter, diaxnater belts, bark calibrators, etc.).
Camera and equipment (wide angle lens).
Tape recorder (batteries).
Shovel.
C - Sample collection
Plant pressing equipment.
Packing material (local newspapers can be used).
Plastic bags.
Bottles for liquid samples.
Preserving fluids.
Growth corer (for wood samples).
Carpenter's brace and drill (for resin samples).
Insulated container (e.g., ice box).
Magnifying glass.
Insecticide spray (for herbarium mater ial).
ALSO: Any needed medicines, camping equipment, vehicles and equipment as needed.
I/ From Burley and Wood (1979)
-58-
Annex 2
SAMPLE .SEED COLLECTI ON DATA SHEET
FAO/DANIDA SEED CENTRE, HUMLEBAEK
DAN/FAO NUM.
SEED COLLECTION DATA DAN/FAQ TREE SEED CENTRE
Scientific name: Provisional number:
Latitude :
Longitude :
n
E Altitude :
A
Map reference:
Country:
Province :
Region and/or administrative unit:
S
I
T
E
S
T
A
N
D
Soil type:
Slope:
Drainage :
Monthly rainfall distribution:
uneven
Plant association:
Density: open
Height :
Diameter:
Stand condition:
Method:
Number of trees:
E
C Amount of seed/cones:
Orientation:
Annual rainfall:
dense Regeneration method:
Age:
Bole:
Date of collection:
Spacing of trees:
Condition of seed/cones:
Potential for commercial scale collection:
N
S Extraction method:
E
E
D
Yield per unit of volume:
Description written
Treatment :
Germination:
Collector:
- 59 -
Annex 3
SAMPLE SEED COLLECTION SHEET
COMMONWEALTH FORESTRY INSTITUTE, OXFORD
Species: Pinus oocarpa Shiede Seed No.: K31 Storehouse No. 1/71
Country: Nicaragua Province: Nueva Segovia
District: Dipilto Area: El Junquillo
Latitude: 13%2'N Longitude: 8635 ! W Altitude: 1000m
Situation; In western Nicaragua on the southern slopes of the Dipilto Cordillera. The
Cordillera forms the northern boundary of Nicaragua in this area. The stands, which are
some 5 km north of Macuelizo and 8 tan west of Dipilto, are part of an area of some 150 000
ha of open pine forest which extends more than 70 km east and west along the cordillera
and northwards to the pine forest of Honduras, Rainfall gradually increases eastwards and
in this direction the lower slopes (below 800 m) gradually fill with populations of P.
caribaea, whereas on slopes above 1 500 m P. pseudostrobus and deciduous forests tend to
dominate. On the astern side near Macuelizo, the low slopes and valleys support only
dry thorn bush, and ?. oocarpa is the only pine species. This is the major pine species
throughout the cordillera.
Soil; Sandy, quick-draining soils or pebbly soils with abundant quartz, derived in situ
by wearing of granite rocks, with frequent outcroppings on the steepest slopes and peaks.
Active erosion. Soils are usually shallow except in depressions or valleys, and contain
hardly any humus. pH = 5-7
Climate: Mean annual rainfall in Macuelizo (5 km to the south) is 90^ mm with the following
distribution (in mm):
MM N D
21 16 113 180 110 85 1U9 163 1*9 13
No temperature data is available for Macuelizo but at the Ocotal weather station (20 km
east and kQQ m lower than El Junquillo), mean monthly maximums in the dry season range
from 28-32C.
Description of stand; Very open pine forest on very steep slopes (25-35) with very scant
herb cover, including Andropogon spp, and Pennisetum spp. Some Quercus spp. are found in
very wet valley areas. The largest pines are over 30 m in height with a diaaneter at
breast height of 80 cm. Pine regeneration is scattered but good in various places. Tree
rings are not very well-defined and are difficult to interpret, but it seems that growth
is slow, from 3-5 rings per cm.
Cone-bearing trees ; DBH UO-60 cm Height : 25-30 m
Branch angle: 70-80 Bole: Straight, cylindrical,
whole
Collection methods; Selected trees on cutting sites (36 trees)
Collection date: January, 1971
- 60 -
STORAGE, TESTING AND CERTIFICATION OF FOREST SEEDS
B. Ditlevsen
National Forestry Service, Denmark
CONTENTS
Page
Introduction 61
Storage 61
Principal factors affecting viability 61
Moisture content 61
Temperature 62
Oxygen and other factors 62
Special factors affecting viability 63
Ripeness of the seed 63
Fungus, bacteria and insects 63
Mechanical damage 63
Cytological and genetic changes 64
Storage methods 64
Seeds of natural longevity 64
Seeds that can be stored 5*10 years with low moisture content and at low
temperature 64
Seeds that can be stored 3-5 years with moderate moisture content and at low
temperature 65
Seeds that can be stored 1-3 years with high moisture content and at low
temperature . . . 66
Very short-lived seeds 66
Tests 67
Sampling 67
Purity analysis 67
Viability tests 67
Germination analysis ... ... 68
Indirect viability tests 69
Results of viability tests and their use 70
Analysis of moisture content 70
Determination of seed weight 70
Health tests 70
- 61 -
Page
Seed testing equipment 71
Certification of forest seeds 71
Certification programmes 71
Implementation of certification programmes 71
Bibliography 73
INTRODUCTION
Effective planning and production of afforestation programmes depends on the avail-
ability at all times of sufficient quantities of seeds with the right physiological and
genetic characteristics. In the first place, the seed must be stored until required,
without losing its germi native capacity. Secondly, there must be continous checking,
by testing the physiological characteristics of the seed. Finally, it is important for
the future production of plantations that the material have the right genetic quality.
STORAGE
Storage may be defined as the conservation of live seed from harvesting to sowing
(Holmes and Buszewicz, 1958).
The reasons for storing forest seeds may be summarized as follows (FAO, 1955):
1) To conserve seeds in the best conditions to protect their germinative capacity
from harvesting to sowing;
2) To protect the seed against destruction by rodents, birds and insects;
3) To preserve quantities of seeds harvested in good years, in order to have
reserves available for years when little or no seed has been produced.
Many forest species only produce seeds in sufficient quantities at intervals of
several years; long- term storage is extremely important for a regular supply of forest
seeds.
It is usually assumed that respiration and metabolic activities must be considerably
reduced if the seed is to survive a long period of storage. In practice, this means that
only seed which can tolerate a reduction in its moisture content, and storage in these
conditions, can survive a long storage period. It is also important that the seed be fully
ripe before drying and that it not suffer damage during harvesting and handling.
PRINCIPAL FACTORS AFFECTING VIABILITY
Viability may be generally defined as the capacity to survive or continue develop-
ing. A live seed is one that is capable of germinating in favourable conditions. The
germination of the seed is not necessarily either easy or rapid, and dormant live seeds
may require lengthy special treatment to germinate (Owen, 1956). The roost important
factors influencing seed viability during storage are the moisture content of the seed
and the temperature. In some cases it has been shown that various gases in the surrounding
area influence the seed somewhat (Owen, 1956).
MOISTURE CONTENT
According to the rules of the International Seed Testing Association (ISTA), the
moisture content of the seed is expressed as a percentage of the weight of the seed in
- 62 -
its original moist state (ISTA, 1976):
Percentage of moisture = Weight of moisture x 100
Weight of moisture +
Weight of dry matter
A reduction in moisture content considerably slows the metabolic processes, which,
in turn, slows the respiratory process and the consumption of stored nutritional matter.
However, it must be pointed out that some forest species with large seeds - for example,
many leafy species - do not normally survive drying. It is also important that they be
dried with great care.
The moisture content of the seed increases before ripening. After ripening it is
much lower. The natural moisture content in seeds after ripening varies between 15 and 50
percent, depending on the species and on conditions.
A change in moisture content during storage as a consequence of storage outdoors,
or due to the opening and closing of the storage containers, destroys the capacity of the
seed to germinate (Barton, 1961).
The moisture content obtained in storage depends to a large extent on the relative
humidity of the air inside the storage area. After a few days or, in some cases, a few
hours a characteristic balance is reached between the moisture content of the seed and the
relative humidity of the air (Holmes and Buszewicz, 1958). Changes in the moisture content
depend on the relative humidity of the air, the moisture content of the seed and type of
testa and size of the seed lot.
It should be borne in mind that the results of studies of moisture content of a
seed lot are mean values, and that there could be major variations in moisture content
from seed to seed. Schonborn (1964) demonstrated, for example, that in one lot of Pinus
sylvestris seeds with an average moisture content of 6.7 percent, individual percentages
varied from 4.1 to 9.3.
TEMPERATURE
Seeds usually keep better at relatively low temperatures. Variations in temperature
have an adverse effect on seed quality.
The interaction between temperature and seed moisture content is very important in
storage, and it is often difficult to separate the two factors* As a general rule, it may
be said that when the temperature is low, the critical moisture content is higher than when
the temperature is high. In other words, low temperature can to a certain extent compen-
sate for high moisture content, and vice versa (Holmes and Buszewicz, 1985).
Temperature influences absorption of moisture by the seed during storage. A general
rule is that moisture content of the seed at a given relative humidity drops when the
temperature goes down.
Schonborn (1964) made studies to demonstrate the lowest temperatures that could
be tolerated by seeds of different forest species at different moisture contents.
Results show that seed tolerance of low temperatures is proportionate to low
moisture content. Also, it appears from the studies that species such as Pices abies,
Abies alba, Pinus Sylvestris, Fagus sylvatica, Pseaudotsuga, Betula and Quercus tolerate
temperatures as low as -20C at a moisture content between 8 and 10 percent.
For seeds requiring a high moisture content in order to conserve their gerrai native
capacity, storage at temperatures below zero leads to frost damage, with the consequent
loss of germi native capacity (Holmes and Buszewicz, 1958).
OXYGEN AND OTHER FACTORS
The respiratory process depends naturally on the moisture content in the seed. The
purpose of lowering temperature and moisture content during storage is precisely that of
- 63 -
lowering respiratory intensity, since this is the only way of ensuring that the seed main-
tains its germinative capacity for a long time*
It is of course desirable to store seeds with low moisture content in airtight
containers that maintain a constant humidity, slow the respiratory process and protect the
seed against pests. For seeds requiring a high moisture content during storage, such as,
for example, Quercus spp., storage in airtight containers can be harmful, apparently due
to the fact that this group needs some change of air during storage (Wang, 1977). However,
it is clear that air composition within the container undergoes modification with time,
due to respiratory processes, something which could influence the life of the seed,
A series of studies has been made of the influence of different gases on the life of
the seed, and the results have been published by Owen (1956) and Barton (1961).
SPECIAL FACTORS AFFECTING VIABILITY
RIPENESS OF THE SEED
How ripe seed is at harvest is of course an important factor, and is responsible
for some variations in seed viability. The timing of harvesting is therefore highly impor-
tant, and it is also important to be able to determine the earliest stages at which a large
quantity of live seeds can be collected.
The degree of ripeness can be assessed by a series of different methods, described
in detail by Earner (1975).
However, it is often difficult to decide the most appropriate moment for harvesting
seeds, especially in tropical countries where the period from maturation of the seed to
shattering is often very short. In other cases there can be various degrees of ripeness
on the same piece of land, or even the same tree, which can mean that unripe seeds have
to be harvested, in the hope that post-harvest ripening can be induced before storage.
FUNGUS, BACTERIA AND INSECTS
Seeds stored under relatively humid conditions are easily attacked by the so-called
storage fungi, of which the major groups are Aspergillus, Botrytis, Rhizopus and Penicillium.
Holmes and Buszewicz (1958) provide more references.
The most widely used method of avoiding attacks by fungi and bacteria during storage
is the use of relatively low temperatures or relatively low moisture contents, under which
fungi and bacteria cannot survive. The use of both a low temperature (maximum +5 C) and
a low moisture content (maximum 10%) provides the best natural protection.
According to Christensen (1972), the use of fungicides during dry storage often poses
problems, since many fungicides have to be dissolved in water.
It is unusual for seeds which were not already infested before storage to be destroyed
by insects, especially Megastigmus species.
In most cases, such losses can be avoided by controlling the temperature, bearing in
mind that almost all insects in stored seed die at temperatures above 40-42C (Holmes
and Buszewicz, 1958). Chemical products can also be useful in controlling insects; however,
it should be borne in mind that these products can considerably reduce germinative power,
particularly in the case of seeds with a relatively high moisture content (Ezumah, 1976).
MECHANICAL DAMAGE
Mechanical damage can be defined as harmful changes due to damage during harvesting
or handling of seed. Since the effects of the damage increase with time, mechanical damage
also covers heavy damage resulting from mechanical damage (Moore, 1972).
The reaction of seeds to damage varies very much, and some species have a greater
natural capacity for recovery. Major damage will immediately lower seed viability, while
- 64 -
the effects of less important damage will usually only become apparent after a certain time.
Consequently, it is important that seeds which are going to be stored for a long time
should be treated with care during both harvesting and subsequent handling.
The moisture content of the seed is important because of its effect on resistance
to mechanical damage. Moist seed tends to swell, whereas dry seed tends to break.
The best method of detecting the presence and kind of damage is tetrazolium testing
(Moore, 1969). Radiographic methods such as those described by Kamra (1967) also produce
good results.
CYTOLOGICAL AND GENETIC CHANGES
Lowered viability means that yield is reduced in two ways. First, a drop in the
germination rate lowers the number of plants per area unit, and second, the surviving
plants may be of inferior quality (Roberts, 1972) .
Many experts fear that a considerable reduction of germinative power involves
changes in the genetic composition of a seed lot since the survival capacity varies from
one genotype of another (Harrington, 1970, Frankel, 1970). They also fear that poor
storage conditions may increase the number of chromosome changes (Harrington, 1970).
Wang (1977), however, reached the conclusion that these two kinds of genetic modi-
fication can largely be controlled and kept to a minimum by efficient harvesting, handling
and storage of seed.
In practice, this means that seeds, stored under good conditions and which have
maintained their germinative capacity can safely be used. But it is not very safe to
use seeds with severely curtailed germinative capacity, when seeds have been stored for
long periods, they should be used with circumspection, particularly in breeding and gene
conservation.
STORAGE METHODS
As many be seen from the above, there are great differences between different species,
or rather, between different genera, as regards capacity to survive storage. Seeds can be
divided into five major groups in the light of this capacity. Guidelines as regards
storage methods for the different groups, and examples of genera and typical species within
each group, are given below.
1) Seeds of natural longevity. This group included Acacia, Robinia, Albizia,
Sophora, Cere is, Cytisus and Gleditsia, all legumes, whose seeds have a hard
testa and normally a very l ow moisture content.
The seed should be stored in a dry place; the temperature, however, is not very
important, and it is not necessary to use airtight containers. In some cases it
is necessary to use disinfectants and other control measures.
2) Seeds that can be stored 5-10 years with low moisture content and at low
temperatures. Typical examples of this group are Picea and Pinus .
Preparation for storage: The seed should be handled properly from the beginning,
and should not be harvested until it is fully ripe. The cones should be kept
cool, and damage to the seed and abrupt changes in moisture content and tempera-
ture should be avoided. There should be adequate ventilation during extraction,
and temperatures should not exceed 20 C at the start and 40 C at the end of
extraction. After extraction the moisture content is very often 5-8 percent, and
the seed should be dewed (trimmed) and cleaned as soon as possible before being
stored. At the above-mentioned temperatures it is not normally necessary to
disinfect the seed.
Drying process* The moisture content of the seed should be ascertained as soon
as possible, by one the methods described in the paragraph on seed analysis.
- 65 -
If the seed does not have the right moisture content, it must be dried. Drying
can be in the sun, in hot chambers or in an oven. It is important that air
circulate freely during the drying process (Wakely, 1954), and that the tempera-
ture not exceed 30C. Small lots can be dried chemically with CaO, sulphuric
acid or CaCl 2 (Magini, 1962).
Moisture content and storage temperature. For most species of the genera
Larix, Picea, Pinus, Pseudotsuga, Thuja, Tsuga, Chamacyparis, Cupressus,
Cryptomeria, AInus , Betula and Eucalyptus , the following can be recommended
(Schonborn, 1964, Wang, 1974 and Earner, 1975):
Storage period
Moisture content %
Temperature C
3-5 years
More than 5 years
More than 10 years
6 - 8 %
6 - 8 %
6 - 8 %
+2 to +4C
-10 to +4C
-10C
In the case of Eucalyptus, Barner (1975) states that most species can be stored
for a period of up to ten years in airtight containers at temperatures of 1-5 C,
if the moisture content is kept at 4-8 percent. Turnball (1975), Filho and
Lisbao (1973) provide more detailed information on some species of Eucalyptus.
As mentioned above, moisture content and temperature are closely linked. If
factors fall short of the optimum, it can be offset to a certain extent by the
other.
Storage containers. Cold storage rooms where the temperature can be kept at the
desired level are normally very expensive to build, so it is important to store
the seed in airtight containers. Glass or metal containers were originally
used for this purpose. More recently, plastic containers and polythene bags
have begun to be used; however, they are containers not recommended for long-
term storage, since they are not entirely damp-proof (Owen, 1956).
Seeds that can be stored 3-5 years with medium moisture content and at low
temperature.
a. This group includes Abies (Barner, 1975) and Cedrus and Libocedrus
(Holmes and Buszewicz, 1958).
Preparation for storage. Heat should not be used in extracting the seeds.
The cones should be stored in a well-ventilated place until they begin to
separate, after which a mechanical handler can be used, very carefully, to
activate extraction. Care should also be taken in later handling, since
seeds are extremely sensitive.
Drying for storage. Abies seeds should not be dried until a 1-2 month
after ripening period. Drying should be done carefully at temperatures
not exceeding 25 C.
Moisture content and storage temperature.
Temperature C
Storage period
Moisture content %
1-3 years
More than 3 years
12-13 %
7-9 %
-4to-15C
-10 to-20C
Storage containers. See group 2.
b. To this group belong the leafy species, Acer, Fagus , Fraxinus, Ulmus,
and possibly Tectona (Barner, 1975).
Preparation for storage. With the exception of Ulmus, whose seeds should
be harvested before maturity, seeds should be fully ripe before harvesting.
- 66 -
The seed should always be spread out in well-ventilated rooms and turned
over at regular interval.
Drying for storage. Normally it is not necessary or desirable to dry
seed artificially.
It should be pointed out that Fagus sylvatica seed should after-ripen
under cold, humid conditions for another two or three months , after
which the moisture content can be carefully reduced (Barner, 1975).
Moisture content and storage temperature. For most species in the group
the following may be recommended:
Storage period
Moisture content %
Temperature C
1-2 years
2-3 years
More than 3 years
20 - 25 %
12 - 20 %
7 - 10 %
-4C
-4 to -10C
-10C
For Tectona grand is there are no precise data, but it seems that the seed
can remain viable for 2 to 3 years if it is in dry storage in plastic bags
(Murthy, 1973).
Bonner (1978) estimates that the storage period for Triplochiton scleroxylen
and Gmelina arborea is the same; seed is stored at temperatures of - 5 C
and moisture contents of 5 - 10 percent.
Storage containers. See paragrah 2. However, it should be observed that
seeds with a moisture content of more than 15-20 percent should not be
stored in airtight containers, but in bags which allow some air to circulate.
4) Seeds that can be stored 1-3 years with high moisture content and at low tempera^
ture.
This group includes various leafy species with large seeds and fruits, such as,
Aesculus, Castanea, Juglans, Liriodendron and Quercus (Barner, 1975).
The most important principles of storage are good ventilation, uniformly high
moisture content, and moderate to low temperature.
Holmes and Buszewicz (1958) provide a series of examples of methods of short-term
storage (in the winter), both in the open air and in a building or under a
protective covering. As a general rule, these methods are not used for long-term
storage; also, it is very difficult to control moisture content and temperature
during storage.
5) Very short-lived seeds
This group includes Salix and Populus (Barner, 1975).
Normally the seed loses its viability very rapidly, but with some species the
seed can be preserved longer (Jones, 1962). Holmes and Buszewicz (1958) and
Wang (1974) have provided more detailed information on individual species*
TESTS
The purpose of seed analysis is to obtain actual data on a given seed lot
(Justice, 1972).
One of the most important factors in successful testing is the use of standard
methods providing comparable, replicable results. In this connection, mention
may be made of the International Seed Testing Association (ISTA), which prepares
and publishes standardized methods for seed testing. The last updated edition of
ISTA standards was published in 1976 (ISTA, 1976).
- 67 -
SAMPLING
For information from a seed lot to be reliable and accurate, the sample taken (and
this is the purpose of the analysis), should be representative fo the seed lot. It is
therefore important to use standarized, objective sampling methods. Even very precise
analyses are less valid if the sample taken is not representative.
Probes are often used, reaching all parts of the bag and designed in such a way
that they extract samples of an equal volume.
Normally, the sample taken should be reduced to a standard portion, and also at this
point it is important that the reduction should be done objectively and correctly. Turnbull
(1975) describes various mechanical and non-mechanical methods of reducing the sample.
PURITY ANALYSIS
Forest seeds can contain impurities in the form of weed seeds, seeds of other tree
species, bits of seeds, leaves and other things. The purpose of seed analysis is to make a
quantitative assessment of the composition of the lot under analysis. To this end, the
lot is divided into its component parts (Turnbull, 1975).
Pure seed means seeds of one species. As well as ripe seeds, these can be under-
sized, shrivelled, immature, or germinated seeds, providing they can be identified with
certainty as belonging to the species in question.
Parts of seeds more than half the original size (ISTA, 1976) are also included.
Leguminosae and Coniferae seeds without testae belong to the inert matter group.
Inert matter includes seedlike structures, for example, broken wings and conifer
seed coats.
Foreign elements include sand, stones, leaves, bark and any other element which is
not seed.
The entire sample is weighed, including all impurities, after which the pure seed is
separated out and weighed separately.
Calculation of the percentage of purity is expressed in the foiling way:
Purity (7.) = Weight of pure seed 1
X 1 UU
Total weight of sample
Most of the work of dividing the lot to be analysed into the relevant component
parts has to be done manually, using a magnifying glass, stereomicroscope, etc. However,
there are today various apparatuses and methods of reducing the manual work to a certain
extent. First, there is the seed blower, with which it is possible to divide the lot into
heavy and light componenents. Among other methods may be mentioned: separation by density
(Stermer, 1964), use of a vibrating table (Guldager, 1973), and X-rays (Kamra, 1965).
The pure seed component is often used for germination analysis, and is therefore
important in' any evaluation of the productive potential of a seed lot, provided that purity
and germination analysis are both tested.
VIABILITY TESTS
It is difficult to give a definition of seed viability broad enough to be applica-
ble in all circumstances. For roost purposes, a seed is considered viable if it germinates
under favourable conditions, assuming potential dormancy has been eliminated (Roberts, 1972).
Methods of ascertaining germination potential and methods of indirectly measuring
seed viability are described overleaf.
- 68 -
GERMINATION ANALYSIS
The fundamental purpose of any laboratory analysis of germination is to estimate the
maximum number of seeds that would germinate under optimum conditions.
Germination is expressed as the percentage of pure seed producing normal plants, or
as the number of seeds germinating per weight unit of the seed lot. It is advisable to
make germination analyses in accordance with 1STA standards.
PREPARATION OF THE SAMPLE FOR GERMINATION ANALYSIS
All germination analyses should be made on pure seed.
The seed should be mixed and counted thoroughly, and the analysis should be repeated
several times. For more information on details concerning preparation and use of laboratory
equipment, see Turnbull (1975).
GERMINATION EQUIPMENT
The choice of germination equipment will depend both on the quantity and the type
of seed. The most important equipment recommended by ISTA is briefly described below:
1) Jacobsen and Rodewald apparatus. The seeds are placed directly over a water
bath or damp sand. The substratum or filter paper bed is kept moist by means of
a wick from the substratum down to the water. Alternatively, porous dishes
containing seed can be placed directly over the damp sand or over water. The
temperature of the container is controlled automatically.
2) Germination cabinet. In this apparatus trays can be fitted in layers, therefore
requiring less space. Also, ambient humidity, temperature and light can be
controlled to a certain extent.
3) Room Germinator. If a great many analyses are to be made, whole rooms can be
used with the appropriate equipment to control temperature, humidity and light.
GERMINATION CONDITIONS
Optimum conditions for the different stages of germination are not identical and can
very often vary with different seeds within one seed lot. As a result, most of the research
on seed has aimed at ascertaining the combination of conditions that provides the most
uniform, rapid and complete germination of most samples in any one species.
Turnbull (1975) comments in detail on the following factors which roust be borne in
mind in germination analyses:
1) Germination substratum;
2) Humidity and ventilation;
3) Temperature control;
4) Light;
5) Distance between seeds;
6) Fungus control.
DORMANCY
The seeds of most tree species germinate upon exposure to the right humidity and
temperature. However, seeds of some species do not germinate even under favourable condi-
tions, until they have undergone a physical or physiological change. This state is called
dormancy. The state of dormancy may be transitory, and many species exhibit dormancy only
during the first six months after harvesting.
- 69 -
Turnbull (1975) mentions the following types of dormancy:
1) Embryonic dormancy. "Embryonic dormancy 11 means the situation in which germina-
tion of the fully developed seed appears to be blocked by internal factors.
2) Dormancy of the testa. Many species of seed do not germinate because they have
a thick testa.
3) Induced or secondary dormancy. A few species can enter into a state of dormancy
due to incorrect storage or handling, and this type of dormancy is called secon-
dary dormancy.
4) Unripe embryo. Seeds in which the embryo is not entirely developed when the
fruit is ripe require a period of after-ripening before germination.
5) Mechanical resistance of the testa. In some cases considerable force is needed
to break the testa. This applies to certain species of Prunus and Tectona grandis.
As in the case of dormancy of the testa, scarification can be used to break the
testa.
6) Double dormancy. Seeds of some species have two types of dormancy that have to
be eliminated before they germinate. The most usual combination is embryonic
dormancy and testa dormancy.
EVALUATION OF PLANTS
A seed is said to have germinated when it has developed into a normal plant.
Various groups of abnormal plants (see more details in ISTA, 1976) are not included in
germination trials because they seldom survive. Germination tests (even excluding ab-
normal plants) normally produce over estimates of viability under field conditions, due
to the optimal testing conditions.
INDIRECT VIABILITY TESTS
The purpose of quick viability trials (Turnbull, 1975) is:
1) To ascertain quickly the viability of species that usually germinate slowly
and show signs of dormancy in normal methods of germination;
2) To ascertain the viability of lots which, at the end of the germination test
show a high percentage of hard or ungerminated fresh seeds.
Four different methods are described below:
1) Analysis by cutting open. By opening the seed it is possible to study the
endosperm and the embryo directly and to see whether their colour is normal
and if they are developing normally. The method is not very accurate and
often leads to over estimates of germinative capacity.
2) Tetrazolium tests. By this method, living cells turn red when the colourless
tetrazolium salt is reduced. The procedure has been described in detail by
ISTA (1976).
Justice (1972) points out that the use of this method is in practice limited
by a series of problems, such as the difficulties in staining some seeds;
some seeds have to be opened for the colour to become visible. In several
cases, results are not in line with those of germination tests, and inter-
pretation of the different shades is difficult.
3) Trials of extracted embryos. Extracted embryos are placed on damp filter
paper in Petri dishes. In a few days it is possible to see which embryos are
alive and which are not.
- 70 -
4) X-ray technique. This method reveals empty seeds, mechanical damage, abnormal
development of the internal structure of the seeds, and the thickness of the
testa. It also assesses viability, using a technique of staining or contrasting
matter.
RESULTS OF VIABILITY TESTS AND THEIR USE
The results of the viability tests may be described as follows (Turnbull, 1975):
1) Percentage of germination. This is the percentage of seeds in the lot which
have germinated by the end of the trial.
2) Germi native potential. This is the sum of the germinated seeds and the remaining
sound ungerminated seeds.
3) Germination energy. This indicates the percentage germinated within a given
time, for example, 70 percent within seven days, 90 percent within 14 days, etc.
4) Number of live seeds per weight unit. The indication could be useful to the
people who are going to use the seed.
ANALYSIS OF MOISTURE CONTENT
As mentioned above, determination of moisture content is very important for storage
purposes.
1STA recommends the following three methods of determinating moisture content (1STA,
1976):
1) Oven drying method (130 C). Owing to the high temperature, this method cannot,
however, be used for forest seeds.
2) Oven drying method (105 C).
3) Toluene distillation method. This method should be used to determine the moisture
content in seeds containing volatile oils, for example, Abies spp.
As well as the methods mentioned above, there is a series of automatic measurements
of moisture content. They are not sufficiently accurate for official trials, but they can
provide rapid and fairly reliable information on the moisture content of seeds.
DETERMINING SEED WEIGHT
Determining weight of 1 000 seeds has been described by 1STA (1976).
HEALTH TESTS
Health tests are run to determine the presence of micro-organisms or diseases in
seeds. These trials are important for three reasons:
1) Disease-carrying seeds can lead to forest outbreaks of pests, lowering the
commercial value of the trees.
2) Through seed lots, diseases can spread to new regions. Quarantine investigation
and certification for the international trade may therefore be necessary.
3) Health tests provide more data on the causes of abnormal seddlings, and could
supplement germination analysis.
The 1STA standards provide general rules for health tests.
- 71 -
SEED TESTING EQUIPMENT
Both the seed testing equipment and its use are described by ISTA (1976).
CERTIFICATION OF FOREST SEEDS
The purpose of certification of forest seeds and plants is to preserve and make
available to working foresters sources of seeds, plants and other propagating material
of superior provenances and cultivars that have been cultivated and distributed, thus
ensuring the genetic identity and superior quality of seeds and plants (Matthews, 1964).
CERTIFICATION PROGRAMMES
Barber (1969) summarized the problems relating to control of genetic identity in the
following way:
Good results in tree improvement and afforestation programmes depend on accurate
control of the genetic identity of reproductive material. We should try to use only
reproductive material of known genetic identity. However, the precision with which we can
identify the material will vary according to species, locality and final use.
The forest tree breeder must have full knowledge of the source of the seed plasm
with which he is working. Tt is particularly important that the identity of each tree be
maintained, so that the researcher may consider the risks of any negative trait from the
mating of related individuals. As offspring are produced and grow, the breeder needs to be
able to study the seneolofcical record of each individual to pinpoint which parent has
contributed which character, whether desirable or not. He/she should be in a position to
duplicate all crossings if necessary, and must catalogue and identify accurately all the
material exchanged or approved for use.
The forester should know which source or line best meets his needs. To achieve
optimal results, he/she should know the exact source of the material used for the estab-
lishment of plantations and for stand regeneration; knowledge of the genetic identity
of the material used is necessary to plan proper spacing and cultivation techniques. As
the use of a disease-resistant strain, for example, lowers mortality and produces fewer
defects, the forester can space trees further apart or thin more frequently. Reports of
major genotypic interaction in the environment indicate that tree breeders can develop
cultivars that react favourably to qualitative difference in site or cultivation methods.
If seeds or seedlings are not available from the appropriate source, it could be desirable,
from the economic point of view, to postpone planting for a year or more.
For a certification programme to function satisfactorily, there should be a mechanism
to facilitate the participation of all interested parties in the formulation of operational
procedures. Certification should be backed by legislation.
Several countries use certification programmes (Matthews, 1964). Some programmes,
such as the North American programme, are regional.
An international programme under OECD was established in 1967. In 1974 a few
modified rules were prepared (OECD, 1974).
The OECD programme in based on the voluntary participation of member countries, but
NATO member countries can also participate.
IMPLEMENTATION OF CERTIFICATION PROGRAMMES
A broad programme should include the following elements:
Planning
1) Preparation of maps, with an indication of the distribution of important species.
2) Delimitation of regions of provenance of these species.
- 72 -
3) Delimitation of important regions for afforestation and reforestation.
4) Estimate of supply and demand of seeds and plants.
Execution
5) Organization and administration.
6) Classification and approval of sources.
7) Recommendations for the choice of provenances and transfer of reproductive
material.
8) Production and control measures.
9) Record of data and documentation.
10) Sale of reproductive material.
Ref. items 1-A. These items are important for evaluation of the need to establish a
certification programme. An estimate of the quantity of seeds and plants necessary should
be made and compared with available sources, both on the site and elsewhere.
Ref. item 5. This item includes the designation of the various groups of authorities
responsible, such as:
1) a director or directors;
2) an advisory group;
3) a working group for approval of sources;
4) inspectors.
Ref. item 6. This item is fundamental to the certification programme, and has been
discussed in detail by Barner (1973). Major sources could be listed as follows:
1) Regions of provenance;
2) Stands;
3) Regions of seed production;
4) Individual trees;
5) Seed plots.
Approval of sources, in accordance with certain minimum requirements, would be the respon-
sibility of the above-mentioned working group for approval of sources.
A national list of sources should be prepared. The list should include the most important
data, such as Latin name, identification, location, origin, ecological conditions and
results of trials. A definition of classes and instructions for collection of information
has been provided by Barner (1975). It is not necessary that the national list include all
the above-mentioned types of source, but it should be possible to extend the programme to
include all types later.
Ref* item 7. One later development of the approval of sources could be the recommendation
of these sources for general or specific uses. Sources have often been approved on the
basis of the phenotypical appearance of the seed trees or on the basis of very limited
progeny texts. Recommendations should be based on tests, but since these take a long time
and seed must be produced, it is necessary to use sources of seed that have not yet been
tested. In certification programmes, therefore, a distinction is made between tested
material and untested material within each class of the approved sources. The certification
- 73 -
programme should be established and developed in close collaboration with forestry research
and particularly with forest tree breeding.
Ref . item 8. As an example of rules relating to production and control measures, Barner
(1975) indicates the minimum standards and requirements of the OECD programme as regards
material from approved sources. The standards establish how inspection is to be carried
out in the forest, the plantation and the nursery, and how to mark and package seeds.
Ref. item 9. Recording data and documentation is an integral part of the control procedures.
Each seed lot, for example, should be accompanied by standard data with detailed indications
of identification, harvesting, seed extraction storage and quality.
Ref. item 10. Once all the operations have been carried out in accordance with the programme,
a certificate can be granted, to which reference can be made when material from the source
in question is sold.
BIBLIOGRAPHY
Barber, J., 1969. Control of Genetic Identify of Forest Reproductive Material. Second
World Consult, on For. Tree Breeding, Washington, D.C. Vol. 2, 11/3.
Barner, H., 1973, Classification of Sources for Procurement of Forest Reproductive
Material. FAO/DANIDA Training Course on Forest Tree Improvment. Limuru, Kenya.
Barner, H., 1975. The Storage of Tree Seed, FAO/DANIDA Training Course on Forest Seed
Collection and Handling, Chiang Mai, Thailand, FAO 1975.
Barton, L.V., 1961. Seed Preservation and Longevity. Leonard Hill, London.
Bonner, F.T., 1978. Storage of Hardwood Seeds, Forest Genetic Resources Information
no. 7, Forestry Occasional paper 1978/1, FAO, Rome.
Christensen, C.M., 1972. Microflora and seed deterioration. In: Viability of seeds by
E.H. Roberts. Syracuse University Press.
Ezumah, B.S., 1976. Seed Handling and Storage, Savanna Afforestation in Africa, Kaduna,
Nigeria, FAO 1976.
FAO, 1955. Handling forest tree seed. FAO, Rome
Filho, W.S. & Lisbao, L., 1973. Influence of the relative humidity on the characteristics
of the seeds of Eucalyptus saligna. IUFRO Internat. Symposium on Seed Processing, Bergen
Norway 1973. Publ. The" Royal College of Forestry, Stockholm.
Frankel, O.H., 1970. Genetic Conservation in Perspective, Genetic Resources in Plants -
their Exploration and Conservation, 1970.
Guldager, P., 1973. Seed problems related to direct sowing in pots, "Seed Processing"
Proc. IUFRO WKG. Group on Seed Problems, Bergen. Vol.11, Paper 12.
Harrington, I.F., 1970. Seed and Pollen Storage for Conservation of Plant Gene Resources,
Genetic Resources in Plants - Their Exploration and Conservation. 1970.
Holmes, G.D. & Buszewicz, G., 1958. The storage of seed of temperate forest tree species.
Forestry Abstracts, Leading Article, for. Abs. Vol. 19. Nos. 3 and 4.
ISTA, 1976. International Rules for Seed Testing, Rules 1976, Seed Science & Technology 4.
Jones, LeRoy, 1962. Recommendations for successful storage of tree seed. Tree Planters
Notes. For. Service U.S. Dept. of Agric. No. 55.
Justice, O.L, , 1972. Essentials of seed testing, "Seed Biology", III, (Ed. T.T.Kozlowski).
Academic Press Inc. New York. 361-370.
- 74 -
Kamra, S.K.; 1965. The use of X-ray radiography for studying seed quality in grasses.
Proc. Int. Seed Test. Ass. 30(3): 519-524.
Kamra, S.K., 1967. Detection of mechanical damage and internal insects in seed by X-ray
radiography. Svensk Bot. Tidskr. 61, Stockholm.
Magini, ., 1962. Forest seed handling, equipment and procedures. Unasylva, Vol. 16,
FAO Rome.
Matthews, J.D., 1964. Seed Production and Seed Certification. Unasyla, Vol. 18. (2-3).
Moore, R.P., 1969. History supporting tetra2olium seed testing. Proc. Int. Seed Test.
Ass. 34.
Moore, R.P., 1972. Effects of mechanical injuries on viability. Viability of seeds by
E.H. Roberts. Syracuse University Press.
Murthy, A.V.R.G. 1973. Krishna. Problems of teak seed. 1. Flower and fruit studies.
2. Germination studies. Intern. Symposium on Seed Processing, Vol. II, Bergen 1973.
The Royal Coll. of For., Stockholm.
OECD, 1974. OECD Scheme for the Control of Forest Reproductive Material Moving in Inter-
national Trade. OECD, Paris.
Owen, E. Biasutti, 1956. The storage of seeds for maintenance of viability. Commonwealth
Agri . Bureaux, Bull. No. 43, Farnham, England.
Roberts, E.H., 1972. Viability of seeds. (Ed. E.H. Roberts), Chapman and Hall, London.
448 pp.
Schonborn, A. von., 1964. Die Aufbewarung des Saatgutes dar Waldbaume. Bayrischer
Landwirtschaftsverlag, Munchen. W. Germany.
Stermer, R.A., 1964. Purity analyses of certain grass seeds by flotation techniques.
Proc. Ass. Off. Seed Anal. 54: 73-81.
Turnbull, J.W., 1975. The handling and storage of eucalypt seed. FAO/DANIDA Training
Course on Forest Seed Collection and Handling. Chiang Mai, Thailand, FAO, 1975.
Wakeley, P.C., 1954. Planting the Southern Pines. Forest Service. U.S. Dept. of Agric.
Washington D.C.
Wang, B.S.P., 1974. Tree seed storage. Dept. of Environment, Can. For. Service, Pub.
No. 1335, Ottawa.
Wang, B.S.P., 1977. Procurement, Handling and Storage of Tree Seed for Genetic Research.
Third World Consultation on Forest Tree Breeding, Canberra, Australia 1977.
- 75 -
EXPERIMENTAL DESIGNS
B. Ditlevsen
National Forestry Service, Denmark
CONTENTS
Page
Introduction 76
Requirements of a good experiment 76
No systematic errors 76
Precision 76
Field of application 77
Simplicity 77
Estimate of uncertainty factor , 77
Principles of experimental designs 78
Null hypothesis 78
Purpose of the test 78
Experimental errors 78
Repetitions 78
Randomization 79
Local control 79
Considerations of a practical nature 80
Designs 80
Complete randomization 80
Random block design 8l
Latin square design 82
Incomplete block design 83
Split plot design 8U
Bibliography 85
- 76 -
INTRODUCTION
The principles governing experimental designs may be expressed in the following way
(Brown et. al. , 1977):
1) Observation of a phenomenon.
2) Formulation of a hypothesis on the phenomenon.
3) Testing the hypothesis.
h) Application of Findings.
Tests are tools to test established hypotheses.
Successful test design and implementation are most important factors underlying
the soundness of test conclusions (LeClerg, 1967)-
Forest tree improvement tests have the following main objectives:
1) To evaluate the material for genetic improvement with respect to a series of
desired properties, based on an established hypothesis. In most cases, these
are field trials, where the material is tested under conditions similar to the
growing conditions (habitat conditions) of the improved material. However,
they could also be nursery, greenhouse or laboratory trials.
2) To study the fundamental genetic parameters which are important for research
and for future development in the field of forest tree improvement.
3) TO analyse different treatments, methods of genetic improvement, etc. Like the
studies mentioned under point 2, these analyses should be made initially with a
view to further research in this field.
Irrespective of the purpose of the test, all the different practical factors should
be taken into account in planning it, so that maximum information is obtained.
REQUIREMENTS OF A GOOD EXPERIMENT
Cox (1958) has established the following five considerations in planning a test.
No systematic errors
This means that experimental units containing a particular material, for example a
particular provenance, should not deviate in any systematic way from the other provenances
in the test.
Precision
If there are no systematic errors, the probable magnitude of random error in the
estimate of the effect of the treatment (material), can generally be measured through
standard error. Standard error (SE) is calculated as follows:
(x-x) 2 s
SE ;
n-1 VT
B = standard deviations X * individual observation
X = the estimated mean of X values n * number of observations
- 77 -
The precision of a test will depend on the following factors:
a. The real variability of the experimental material and the accuracy with which the
experimental work has been done.
b. The number of experimental units (and the number of observations repeated per
unit).
c. Experimental design (and the method of analysis, if the latter is not completely
effective).
A general requisite of precision is that the standard error be sufficiently small to
allow us to draw convincing conclusions; on the other hand, it should not be too small. If
the standard error is great, the test will be almost valueless, whereas an unnecessarily
small standard error implies a loss of experimental material.
Field of application
Estimating the difference between two materials in a test, we reach conclusions that
refer to the special set of units used and to the conditions governing the test. If we
wish to apply the results of the test to other conditions or units, another uncertainty is
added, as well as the uncertainty measurable as standard error. It is important that experi-
mental material be tested under conditions not essentially different from those of the
sites where it is planned to use the material.
Simplicity
This is a very important factor which should always be considered in planning a test.
By simplicity is meant not only a simple design, but also that it is desirable to use simple
methods of analysis. Fortunately, in most cases, efficiency in design is matched by simpli-
city in methods of analysis.
Estimate of uncertainty factor
It is desirable to estimate the uncertainty factor of the assumed differences between
materials. This usually means estimating the standard error for these differences, and then,
on this basis, calculating the limits of error (limits of reliability) for real differences
at a given level of probability. From this, the statistical significance of the difference
between two materials can be measured.
The standard error of the difference is calculated as follows:
SE -
n. and n 2 s number of observations on which the two estimated means are based.
The limits of reliability for real differences may be calculated as follows:
XI and X2 the two estimated averages
t value of student's t at a given level of probability
n. and n 2 * number of observations on which XI and X2 are based.
- 78 -
PRINCIPLES OF EXPERIMENTAL DESIGNS
Series of fundamental and general factors in experimental designs is described
below*
Null hypothesis
As mentioned in the introduction, the purpose of the tests is to test a hypothesis
based on existing differences between materials or treatments. Generally, a so-called null
hypothesis is established, which assumes that any possible differences can only be attribu-
ted to contingencies. This null hypothesis is used in combination with a test of signifi-
cance as an alternative to the established hypothesis being studied.
The concept of null hypothesis is very important in experimental designs. When we
set up the test, we know some factors that produce differences between the units of the
test, but there will always be differences not attibutable to specific factors. To use the
null hypothesis as an alternative hypothesis, there must be random distribution of causes
of variation that cannot be attributed to specific factors.
Purpose of the test
It is important to define the purpose of the test. For example, the purpose could
be the study of a series of consignments of seed for subsequent selection of the best con-
signments of seed for further genetic improvement. In many situations the test are multi-
factorial, and in such situations it is essential that the purpose be formulated before the
design is chosen. For example, it would be desirable to know whether the purpose is the
study of one or more of the principal factors of the test, or whether it is rather an
estimate of a specific interaction.
It is also important to formulate the field of validity to which the results will
be applied.
Experimental errors
Experimental errors were described by Fischer (1951) as the lack of uniform gains
in plots which had, on the whole, received the same treatment. The experimental errors that
occur in nursery and field tests are, in the first place, caused by differences in soil
conditions. Previously attempts were made to reduce this error by holding tests under
homogeneous soil conditions using an improved experimental technique. However, experimen-
tal error cannot be completely eliminated, not even under apparently identical conditions.
Instead, attempts are made to measure the error, so that the uncertainly affecting the
conclusions drawn from the data material can be estimated. This then determines the
selection of an experimental design in a given situation. A good experiment is described
by LeClerg (1967) as an experiment that estimates the magnitude of uncontrolled experi-
mental error.
Fischer (1931) formulated the following three decisive principles for obtaining
exact data - principles which form the basis of the experimental design to be established
and of the subsequent statistical analysis:
Repetitions* There are usually two possible ways of reducing error in the estimate
of an effect. One consists in improving the experimental technique, for example, by using
more precise measuring instruments and taking greater care in the collection of data, etc.
The other possibility and, in many cases, the only one practicable, is to repeat the test
a number of times and use the average of the results obtained. Unfortunately, the second
method is not very efficient, since the error of the average is only reduced by the square
of the number of repetitions. Thus the average of four repetitions will only reduce the
error of one-half ( |/7J of the error of the individual repetition. As well as reducing
the error of the estimate, repetitions make it possible to estimate the uncertainty factor
in the experimental results, e.g. calculating intervals of reliability. Repeated measure-
ments in the same plot are not considered as repetitions but can only be considered as
sub-samples, and the variation between them constitutes the sampling error.
- 79 -
Randomization. In any experimental design it is very important, according to Fischer
(1926), that the different units of the experiment should be randomized. Fischer (1926)
considers that "the estimate 11 of experimental error obtained through repetitions depends
on existing differences between plots treated in the same way, An estimate of experimen-
tal error will only be valid for its purpose if, in the arrangement of the plots, we en-
sure that pairs of plots that have received the same treatment are neither closer to-
gether nor further apart nor in any other related way different from plots that have re-
ceived different treatments".
In most cases, randomization is applied to the experimental area; however, it can
also be applied to time.
In practice, randomization can be done in several different ways. In tests of limited
scope, it can, for example, be done by consecutively numbered cards, which are shuffled and
picked at random. Depending on the design in question, randomization can apply to the complete
test or separately to individual lines, columns, etc. This is described in greater detail
for individual designs. If the tests are broader in scope, randomization tables (or, where
a computer is available, standard randomization programmes) can be used.
Local control. Local control means certain restrictions on the random distribution
of treatments or materials in a repeated test, so that the part of total variation not
applicable to comparisons of treatments or materials may be eliminated. Local control
includes the arrangements of blocks and plots in the test. By arranging the blocks in such
a way that the variation of, for example, soil conditions within the blocks, is minimized,
while a considerable variation can be found between the blocks, an important improvement
will be achieved in the precision of the test, since the variation between blocks can be
calculated and eliminated when the data is analyzed.
Plot size and shape, plant spacing, etc., can often vary from one place to another,
depending on the experimental area, the purpose of the test, the volume and quality of the
material, and on other factors often of a practical nature. Concerning plot size and shape,
there is a series of studies which illustrate optimal plot size and shape (Johnstone and
Samuel, 1974). It does not seem possible to set any general standards for the arrangement
of the plots, but the following points are important when choosing both plot size and shape:
1) An increase in plot size will produce a more reliable estimate of the plot mean.
On the other hand, an increase in plot size will lead to a corresponding increase
in the size of the blocks and therefore in the variation within the individual
block. A greater variation within the blocks will reduce the possibilities
of discovering intraplot differences, if any. Thus Wright and Freeland (1959,
1960) observed that a reduction in plot size to single tree plots led to more
effective tests on pine height and diameter.
2) For tests of, for example, different fertilizers, fairly large plots must be
used, perhaps surrounded by buffer strips to avoid any influence from adjacent
parcels. The same applies to tests in which, with time, there could be
competition between trees of adjacent plots.
3) As regards plot shape, in most cases it will be better to select square plots,
thus obtaining the smallest perimeter and therefore minimal influence from
adjacent parcels. However, where there are considerable systematic site
variations, such as sloping land, it may be better to select oblong plots
arranged in such a way that this variation is absorbed within the plots.
As mentioned above, the precision of a test can be improved primarily by increasing
the number of repetitions. On the other hand, a useless increase of precision through a
large number of repetitions will lead to a loss of resources in the form of plants and
labour. It is therefore important to estimate, when preparing the test, the number of
repetitions needed to give the required precision. For example, if it is desired that
a 10 percent difference of materials show as significant at a level of significance of,
for example, 5 percent, the necessary number of repetitions could be calculated via the
- 80 -
following formula:
LSD
LSD = real difference from D
t = Student's t at the desired level of significance (5 percent) and with the same number
of degrees of freedom as the experimental error (S)
S * experimental error
n number of repetitions.
If the approximate magnitude of the experimental error is known, either from previous
tests of the same kind, or on the basis of a general knowledge of the conditions, the
equation could be solved with respect to n, as follows:
t 2 2 S 2
na t . 2 . S
LSD
Consideration of a practical nature
As well as considerations of a theoretical nature with regard to a test, there are
often practical constraints setting up the test.
In the first place, the area available can lead to a series of limitations, which
will oblige the planner to omit several blocks, repetitions or treatments, or to use a
different and often less efficient design. The choice of one or another possibility will
depend on factors related both to the experimental material and to the site. If, for
example, there is a very small variation in soil conditions, the number of repetitions
will naturally be reduced. Another possibility is to reduce the number of treatments or
materials, if it is known beforehand that some are homogeneous and can perhaps be lumped
together in the analyses.
In the second place, there may be problems in setting up the test at several different
sites. If it is desired to apply the results more broadly, it will be necessary to set up
tests at a number of sites representative of the field as a whole.
As well as these factors, it should be borne in mind that some parts of the test
could be damaged or destroyed. If, for example, there is any risk of damage, a relatively
sound design should be selected that could yield valuable information, even if some
values are lacking.
Similarly, it is important that both the design and the methods of analysis be rela-
tively simple, particularly when the people conducting the experimental work are not fully
conversant with more complicated methods.
DESIGNS
With time, various different experimental models have been developed, some based on
the principle of random distribution, others on that of automatic distribution of treatments
or materials in the plots. A series of models is described below, which largely meet the
requirements of a good experiment as described in the preceding section (Jeffers, 1960,
Perarce, 1975).
Complete randomization
In this design there is no grouping of experimental units, and the different treat-
ments are assigned to individual plots completely at random.
- 81 -
The method is very simple and flexible and can be used to advantage under the follow-
ing conditions (Cox, 1958):
1) In very small tests where it is important to have the maximum number of degrees
of freedom in the estimation of error. In a test which contains N plots and a
treatments, there are N-a degrees of freedom, whereas in a test with an equal
distribution of blocks, there will be N-a-b+1, where b is equal to the number
of blocks in the test.
2) In tests where, as far as can be seen, there is no immediate reason for establish-
ing a block design. This could be the case, for example, in laboratory tests
where the environmental conditions are controlled and uniform.
3) In tests where an increase in precision can only be achieved through the use of a
co-variable.
An example of a complete randomization design is illustrated in Figure 1.
A
B
C
C
A
B
C
B
A
B
C
A
B
A
C
B
A
C
18
Figure 1. Complete randomization design with three treatments (A,B,C), each one
repeated six tiroes.
Random block design
In cases where it is hoped that there may be some form of variation in environmental
conditions, complete randomization is not very practical, since the variation between plots
treated in the same way, and therefore the experimental error, will be large, and precision
considerably reduced.
Under such conditions, plots are usually grouped in blocks, so that each block
includes all the different treatments once. With this design, variation between blocks
can be eliminated by statistical analysis, reducing the experimental error and increasing
the precision of the test. Random block designs are simple to establish, and Figure 2
shows an. example of a test that includes four treatments and four blocks. The distribution
of the treatments within each individual block must be random.
Block 1
- 2
- 3
4
A
B
D
C
C
B
A
D
B
A
C
D
D
C
A
B
Random block design with four treatments (A,B 9 C,D) and four blocks.
As mentioned above in connection with local control, blocks should be arranged in
such a manner as to absorb the maximum variation between blocks, while variation should
be maintained within blocks at as low a level as possible. Figure 3 shows examples of a
good arrangement of blocks under the following circumstances:
- 82 -
1* Establishment in an area with a systemetic environmental variation, such as un-
even ground^ near a window in a laboratory, etc.
2. Establishment in an area where there is marked variation in soil conditions; the
plots should be adjoining, irrespective of their physical arrangement, in blocks
within which conditions appear uniform.
Block 1
Block 2
Block 3
Block 4
Systematic
Variation
Block 3
Figure 3. Arrangement of blocks in a random block design under conditions of
1) systematic site variation and, 2) considerable site variation.
The random block design is simple to establish, and analysis and interpretation
of the test are also simple. At the same time, this type of test is very sound, since
the analysis can be made, even if some values are missing, without any major problems
(Brown et al., 1977).
In principle, this design can be used for testing a random number of treatments,
but in practice problems arise to the extent to which the size of the blocks and the
number of treatments are increased. With the increase in size of the blocks, the varia-
tion within the blocks is also increased, and the purpose of establishing blocks cannot
be satisfactorily achieved. In such cases it may be decided to use an incomplete design
(see below) or a reduction in the number of trees per plot.
Latin square design
In this design, plots have been double-grouped, dividing the test into lines and
columns. A specific treatment is made once in each line and column, as illustrated in
Figure 4, From this it may be concluded that the number of plots in the test will be
equal to the square of the number of treatments.
This design is very effective where there are variations in two directions (for
example in greenhouses), but at the same time there are some restrictions with respect
to the number of different treatments that can be studied. For example, ten will require
the establishment of 100 parcels, and four or less will give too few degrees of freedom
- 83 -
to enable a valid test of the treatment to be made.
In this design there are certain restrictions on the possibilities of randomization,
but partial randomization can be used, as described below.
1) One of the formal designs in Fisher and Yates (1957) is chosen at random.
2) The columns are arranged at random.
3) The treatments are designated at random by letters A, B etc., in the formal
plan.
Column
Lines
D
B
E
C
A
A
D
C
B
E
B
E
D
A
C
E
C
A
D
B
C
A
B
E
D
Figure 4. Latin square design with five treatments (A,B,C,D,E).
Incomplete block design
These designs are called incomplete designs; that is to say, the number of treatments
within a block is less than the total number of treatments which are being tested.
In lattice designs the incomplete blocks are grouped so that each group forms a
complete repetition; see Figure 5.
Rep. x
XI X2 X3 X4
Rep. Y
Yl Y2 Y3 Y4
Rep. Z
Zl Z2 Z3 Z4
11
1
8
6
9
8
5
7
12
2
4
7
12
3
9
5
6
11
2
4
6
10
3
1
10
2
7
4
3
1
12
10
8
9
11
5
Figure 5. Rectangular lattice, 3x4, with three repetitions (X,Y,Z).
The number of treatments in a lattice design should be the same as the square
(4x4= 16,' 5 x 5 = 25 etc.) or the product, according to the following formula:
k(k+l) (3 x 4 = 12, 4 x 5 = 20, etc.). These two types are called square lattice design
and rectangular lattice design, respectively. In each case k plots (3,4,5, etc*) are
arranged within each block and k square) or (k+1) (rectangular) blocks within each
repetition. A completely balanced square design will require k+1 repetitions.
It is not possible to establish completely balanced designs for k2 = 36, 100 or 144,
for rectangular designs.
In establishing a lattice design randomization will follow these rules:
1) Use as point of departure a formal design, as illustrated, for example, by
Cochran and Cox (1957). Choose at random the number of repetitions desired
2) Random lie the order of succession of the repetitions*
3) Randomize the order of succession of the incomplete blocks within each repetition.
4) Randomize the plots within each block.
5) Assign at random different treatments to the numbers of treatment in the formal
plan.
Experimental layouts in the field have the same requirements as those described for
random block designs, i.e., that the blocks should be so placed that a variation in the
experimental area is absorbed as an inter-block variation, while the individual blocks are
as homogeneous as possible.
The lattice design is useful, and at least as exact as a random block design, with
the same number of repetitions. However, the statistical calculations involved are
relatively complicated; it could be difficult, especially in the case of missing values,
to make a satisfactory analysis.
Nevertheless, the lattice design has the great advantage that it can always be
analyzed and interpreted as an ordinary random block design.
Split plot design
A split plot design is a special factorial design, inasmuch as it includes two
degrees of repetitions. A set of factors is linked to the larger plots, while the other
set or sets of factors are linked to plots within each one of the larger plots. In other
words, the larger plots are sub-divided into a number of sub-plots; see Figure, 6.
As there are fewer plots than sub-plots, the factors linked to the larger plots and
the sub-plots respectively will be tested with different degrees of precision. It is
therefore an advantage to use this design in situations in which one of the sets of
factors is of special interest, or in situations in which a specific treatment will
require a relatively large area (for example, in fertilizer trials).
The larger plots can be arranged, for example, in random block designs, Latin square
designs or corresponding designs. The allocation of sub-parcels must be random.
Rep.l Rep. 2 Rep. 3 Hep. 4
Figure 6 Design of split plots with two larger plots within each repetition
(indicated by shading or no shading) and three sub-plots within each larger
plot.
- 85 -
BIBLIOGRAPHY
Brown, A.G. & Ma the son, A.C., 1977. Statistics: Design of Experiments. International
Training Course in Forest Tree Breeding, Canberra, Australia.
Cochran, W.G. & Cox, G.M., 1957. Experimental Designs. John Wiley & Sons, Inc., New York,
USA.
Cox, D.R., 1958. Planning of Experiments. John Wiley & Sons, Inc., New York.
Fischer, R.A. , 1926. The Arrangement of Field Experiments. I. Min. Agr. England 33.
Fischer, R.A., 1931. Principles of Plot Experimentation in Relation to the Statistical
Interpretation of the Results, Rothamsted Conference XIII.
Fischer, R.A., 1951. The Design of Experiments, Hafner Publ. Co., New York.
Fischer, R.A. & Yates, F., 1957. Statistical Tables for Biological Agricultural and
Medical Research, Oliver & Boyd.
Jeffers, I.N.R., 1960. Experimental Design and Analysis in Forest Research. Almquist &
Wiksell, Stockholm.
Johnstone, R.G.B. & Samuel, C.J.A., 1974. Experimental Design for Forest Tree Progeny
Tests with particular Reference to Plot Size and Shape, Proceedings IUFRO Joint Meeting,
Stockholm.
LeClerg, E.L., 1967. Significance of Experimental Design in Plant Breeding, Plant Breeding,
R.J. Frey (ed), The Iowa State University Press, Ames, Iowa, USA
Pearce, S.C., 1975. Field Experimentation with Fruit Trees and other Perennial Plants,
Commonwealth Agricultural Bureaux.
Wright, J.W. & Freeland, F.D., 1959. Plot Size in Forest Genetics Research, Pap. Mich.
Acad. Sci., Arts and Letters 44.
Wright, J.W. & Freeland, F.D., 1960. Plot Size and Experimental Efficiency in Forest
Genetics. Research, Mich. State Univ. Agric. Exp. Sta. Techn. Bull. No. 280.
- 86 -
STATISTICAL INTERPRETATION OF TEST RESULTS
B. Ditlevsen
National Forestry Service, Denmark
CONTENTS
Page
introduction . . o /
Fixed and random effect models ... ... . .. . .. . 87
Statistical significance ... ... ... ... ... ... ... ... 87
Tests ... ... ... ... ... ... ... ... tJo
Cni s Qua re test . . ... ... ... ... ... ... ... 07
t test ... ... ... ... ... ... ... 07
f test ... ... ... * ... ... ... ov
Preliminary investigation ... ... ... ... ... ... ... ... 90
Outliers and missing values ... ... ... ... ... ... ... 90
Tests of requisites ... ... ... ... ... ... ... 90
Normality test ... ... ... ... ... ... ... ... ... 90
Homogeneity of variance test .... ... ... ... ... ... ... 91
Additivity test ... ... ... ... ... ... ... ... 91
Transformation ... ... ... ... ... ... ... ... ... 91
Contingency tables ... ... ... ... ... ... ... ... ... 91
Variance analysis ... ... ... ... ... ... ... ... ... 92
Complete randomization ... ... ... ... ... ... ... ... 93
Random block, design .. ... ... ... ... ... ... ... ... 94
Latin square design .. ... ... ... ... ... ... ... ... 95
Incomplete block design ... ... ... ... .. ... ... 96
Split plot design ... ... ... ... ... ... ... ... ... 97
Correlation analysis ... ... ... ... ... ... ... ... ... 98
Regression analysis ... ... ... ... ... ... ... ... ... 98
Covariance analysis ... ... ... ... ... ... ... ... ... 99
Variability components . ... ... ... ... ... ... ... ... 100
Bibliography ... ... ... ... ... ... ... ... ... ... 102
- 87 -
INTRODUCTION
It is important when preparing a test to define the answers one hopes the test will
provide, and the methods of analysis one plans to apply in appraising the results of the
test. Thus, the possibilities of drawing sound conclusions from a test will depend to a
considerable extent on whether or not the test is properly planned. In other words, one
requisite for a good test will be clear formulation of the purpose of the test, the use of
a design that can give the required answers, and careful control of the preparation of
the test.
Using statistical analysis methods, tree breeder can make a statistical interpre-
tation of the test. A knowledge of the experimental design of the material and the
treatments sampled, as well as of the statistical results, are usually sufficient basis
for a later statistical interpretation of the results.
Some general aspects of statistical interpretation are described below, giving
examples of the most widely used statistical methods. Genetic interpretation will be
mentioned only briefly, since this question is discussed in greater detail in another
paper.
FIXED AND RANDOM EFFECT MODELS
In making an analysis of tests, it is essential to define the statistical model to
be used. A distinction is made between so-called "random effect" and "fixed effect" part
(sample) chosen at random from the total of possible treatments. The test results are
applicable to all treatments represented by the treatments sampled.
In a fixed effects model, the purpose is to study a specific set of treatments in
such a way that the test results are not applicable to a group of treatments greater than
those in the test.
As well as the models mentioned above, there are also mixed models, in which one or
more treatments are considered random, while others are considered fixed.
It is important to define the type of model to be used in each individual case,
since both the test of the treatments and the computing of the variability components
will depend on this. Also, the final conclusions and the scope of the test will naturally
depend on these basic considerations.
However, very often it can be difficult to decide if the effects of a test should be
considered random or fixed, and it is difficult to establish univocal standards in this
respect. The problems have been thoroughly discussed by Kempt home (1975), who also gives
examples of how the two different approaches influence the variance analysis test.
STATISTICAL SIGNIFICANCE
The point of departure for the concept of significance is the establishment of a
null hypothesis, and possibly of an alternative hypothesis, on the test material. Roughly
speaking, the null hypothesis is that hypothesis which, as problems appear, it is neither
natural nor advisable to reject, unless observations indicate very clearly that there are
other possibilities.
Significance tests are used to decide whether results obtained from a test are in
conformity with the established hypothesis. These are tests to check whether the results
deviate significantly from the results anticipated under the null hypothesis.
A test size is calculated on the basis of the trial and its theoretical distri-
bution under the null hypothesis is known. If the calculated size of the test is very
unlikely under the theoretical distribution, there is reason to reject the established
null hypothesis* However, the test of significance is not univocal, since the random
variability of the results of the test can mean that some results, through the random
effect, deviate significantly from the results anticipated under the established null
hypothesis.
- 88 -
The principles related to significance tests are illustrated in Figure 1,
2 1/2 %
Significant Not significant Significant
(rejection) (acceptance) (rejection)
Figure 1 Distribution of the size of the test, if the null hypothesis is genuine.
The figure also illustrates a bilateral test at the 5 percent level.
However, on the basis of statistical theory, it is possible to calculate the pro-
bability that a specific deviation will occur at random. Significance tests are there-
fore made at different levels of significance, indicating the probability that the de-
viation could have occured at random. Instead of merely indicating that the results
deviate significantly from the null hypothesis, it is indicated that the deviations are
significant at the 5 percent level.
Finding significance, for example, at the 5 percent level, implies that there will
continue to be a 5 percent probability that the deviation encountered is the result of
contingencies. For this reason, if a test of significance at the 5 percent level is used
to reject the established null hypothesis, there is a 5 percent probability that this
hypothesis has been erroneously rejected, whereas there is only a 1 percent probability
that the rejection of the null hypothesis is erroneous if the deviation is significant
*t the 1 percent level.
A significance test is often considered as an automatic norm related to the "ac-
ceptance" or rejection of the null hypothesis. As mentioned above, a significance test
can never reject with 100 percent certainty a hypothesis, and therefore, the tree breeder
should take these limitations into account, attempting in each situation to supplement
the information from the significance test with information on the volume of data material,
the experience gained from relevant tests, etc. (Brown et al., 1977).
TESTS
In the previous paragraph mention was made of the principles governing significance
tests. We give below a brief description of three tests frequently used, which include a
fairly large number of the most important methods of analysis mentioned in the present
article.
- 89 -
Chi squared test
Chi squared is shown by the following formula:
2 (o. - e.) (o ? - e,) 2 <o n - ej 2
Chi 2 . 1 L + _? 2_ + + __n n_
f e l 6 2
o^ o frequencies observed
e. e * frequencies expected
f number of degrees of freedom
The chi squared test is used to test whether the frequencies observed are in con-
formity with the established expected frequencies (the null hypothesis). The values,
calculated in accordance with the above formula, are adjusted to chi-squared distri-
bution, so they can be tested via theoretical chi squared distribution.
The so-called "t" test is used to evaluate whether the differences between two
means can be considered statistically significant at the given level of probability.
The size of t is calculated according to the following formula:
fc _ 1 2 where
V
n i n 2
x. and x* m estimates of means
S * standard deviations
n, and iu - number of observations forming the basis of x, and x 2
f number of degrees of freedom (O(n.-l) (n 2 -l))
The calculated value is adjusted to the t distribution, which can then be tested
via the theoretical distribution of Student fl s t.
F test
The F test is used to test the homogeneity of two variances. The value of F is
calculated according to the following formula:
F
f l* f 2 * 2 where
S 2
2 2
S. and $2 variances (mean squares)
2 2
f-. and f~ number of degrees of freedom for S. and S~
The test is based on the so-called F distribution (variance ratio distribution).
- 90 -
PRELIMINARY INVESTIGATIONS
Before starting the statistical analysis itself, and interpretation of the re-
sults of tests, certain preliminary investigations should be made, partly to detect
faults or direct errors in the data material, and partly to verify that the requisites
necessary for implementation of statistical analysis have been fulfilled.
Outliers and missing values
In a first study of the data material it should be checked if there are values
that deviate considerably from other data (outliers) and that should be assumed liable
to error. Such investigations can easily be made, for example by tracing each datum
in a diagram, thus localizing and studying in more detail values which deviate marked-
ly, if any. If there are large quantities of data and a central data processor is
available, it might be desirable to use an automatic tracing programme, which is be-
coming standard equirement in all data processing centres.
Another possibility is to calculate estimates of the means (for example, means
of plots) and revise all values which deviate by, say, more than three times the stan-
dard deviation fron the mean total of the test. Since this involves a fairly large
quantity of data, the calculations take a long time, and therefore it is easier to
data process them.
As well as making the above-mentioned tests of individual values, a check should
be made to see whether there are any missing values, and generally, whether the num-
ber of observations per test unit is in conformity with the test plan.
In a simple design like complete randomization, it does not matter, for variance
analysis, if there are one or several missing values, apart from the fact that this
will reduce the number of degrees of freedom for test error. In other designs, it
may be necessary to replace missing values by estimated values, in order to complete
the analysis. These estimates, related to different designs, have been demonstrated
by Freese (1970), Apart from reducing the number of degrees of freedom, missing values
can be a problem in interpretation of results, especially in fairly complicated designs.
Tests of requisites
Most of the usual methods of analysis require that certain requisites be ful-
filled with regard to data material. Tests of three essential requisites (normality,
homogeneity of variance and additivity) are described below.
Normality test
In view of the fact that many statistical methods are based on the assumption
that data material is adjusted to normal distribution, it is often useful to have a
method for the study of this normalcy.
A chi squared test can be used to test for normalcy, showing whether the distri-
bution of the data material can be expected to match normal distribution. Snedecor and
Cochran (1967) give a detailed explanation of the procedure for the test.
The test mentioned can only shew whether there is conformity with normal distri-
bution. If not, special tests can be used to determine Skewness and Kurtosis, where
these exist (Snedecor and Cochran, 1967),
- 91 -
Homogeneity of variance tests
As mentioned above, the F test can be used to test the homogeneity of two vari-
ances. However, if several variances are involved, Bartlett's test (for the homogeneity
of a set of variances) can be used. The test has been explained in detail by Snedecor
and Cochran (1967).
Variance homogeneity is one requisite for the t test. Is is also a requisite for
a correct F test in variance analyses.
Additivity test
In the statistical models established later in the present article, it is assumed
that the individual factors of the model are additive, that is to say, that the
result obtained on one plot or test can be expressed as the sum of the effects of the
model, for example, effects of treatments or blocks and resudual effects. Turkey's
test, described by Snedecor and Cochran (1967), can be used to ascertain additivity.
However, preliminary study of data material will often reveal any lack of addi-
tivity, since in many cases non-additivity will only be revealed by tracing the data
material.
Transformation
If the tests mentioned above show that one or more requisites have' not been fulfilled, ,
it may be necessary to make a transformation of the data material before completing the
statistical analysis. The following are the three most frequently used forms of- transforma-
tion:
1. Arc sine transformation. This transformation can be used for the values of
binomial distribution expressed in percentages in the field from 1 to 100. This
type of data occurs frequently at the nursery stage, when survival, frost damage,
and other such forecasts are made.
2. Square root transformation. This transformation can be used in cases where variance
depends on the mean, for example, in estimating the number of branches by branch
crowns (Bur ley and Wood, 1976).
3. Logarithmic transformation. This transformation can be used in situations where the
variance is proportionate to the mean square. According to Bur ley and Wood (1976),
as in forecasting flowering rates, using a logarithmic scale (1 1-5 flowers,
2 6-15 flowers, 3 - 16-35 flowers, etc.).
A detailed description of the different transformations and their applications is
given by Snedecor and Cochran (1967).
Apart from the arc sine transformation of values expressed in percentages, in most
cases it is not necessary, if the analyses are made on the basis of plot means, to make
transformations of data. However, the individual researcher should study and evaluate in
each situation the consequences that the missing requisites, if any, could have for the
test conclusions.
Contengency tables
In many situations, tree breeders face the problem of having to compare proportions
from two or more independent samples.
Contingency tables are frequently used, especially in relation to qualitative
(Mendelian) genetics, to determine whether a specific variation with respect to a quilita-
tive property is in conformity with what was expected.
- 92 -
However, contingency tables can also be of interest in forest tree improvement,
where quantitative properties are studied in the first place; the tables can be used in
preliminary investigation of the material, or studies of the different treatments of the
material.
As an example, a 3 x 2 table is shown below containing results of an investigation
of two types of planting machines (I and II). The plants are grouped as A * undamaged;
B - damaged, C - felled.
A B C T
observed
Machine I
expected
117
(116.1)
31
(28.6)
7
(10.3)
155
observed
Machine II
expected
131
(131.9)
30
(32.4)
15
(11.7)
176
248
61
22
331
On the basis of the selected grouping (A,B,C,) it can be ascertained, through a
chi squared test, whether a difference is expected between the two types of machine. The
number expected in each square of the table is calculated beforehand as follows:
Expected number of undamaged plants (A) for machine I
155 x 248
331
Expected number of damaged plants (B) for machine I - 155 x 61
"
116.1
etc.
Chi squared is calculated as follows:
(117 - 116. 1) 2 (31 - 28. 6) 2 (15 - 11. 7) 2 , , fl
116.1 * 28.6 "*" 11.7 " i25.
It may be seen from a chi squared table with two degrees of freedom ( (3-1) (2-1) -2)
that the value of 2.38 given there provides no grounds for assuming that the two machines
in the tests are different.
VARIANCE ANALYSIS
Variance analysis is the most widely used method of interpreting the results of
tests. Variance analysis gives the following information:
1) estimates of the relative size or significance of each identifiable source of
variation;
2) estimated differences among the materials, the treatments and the sites used
in the test;
3) indications of the precision of the estimated differences among the estimated
means, through their standard error and their limits of reliability;
4) tests of statistical significance of variances and differences.
The exact form of the variance analysis will depend both on the test design and on
the mathematical model used as a basis. Also, both the significance test and the later
interpretation of the results will depend on whether the individual variables are consi-
dered ma fixed or random.
The essential part of all analyses is, however, to estimate variances linked to each
source of variation (e.g. material, site, residual).
- 93 -
Examples of analyses of five different designs are shown below. (SheffS, 1959.
Searle, 1971).
Complete randomization
In this design there is only one source of systematic variation which can be identi-
fied, i.e., among the treatments. A systematic variation that may exist on the site cannot
be segregated, unless it is contained in the variation from plot to plot.
The mathematical model takes the following form:
*ij ^ * a 4 * EH Where
"13
individual observation
p mean total
1 - effect of the ith treatment
c lj " residual effect
A summary of variance analysis table is shown below.
Source of variation
df
CS
F
Between treatments
a - 1
fV'i--*--' 2
Residual
N - a
1 (x u- x i- )2
Total
N - 1
H -.'
total number of plots in the test
a number of treatments
J i number of plots per ith treatment
Xij individual observation
*i. estimated mean of ith tretmant
*' mean total
The F test appears in the last column of the analysis table. If the F test reveals
significant differences between the treatments or the materials, the real difference of D
can be calculated, and the limits of reliability can also be calculated for the individual
differences among the estimated means.
The real difference of D (LSD) is calculated in accordance with the following
formula.
VI . S
LSD
where
' Vff
t - the value Student's t at a given level of probability,
S test errors (residual MS (mean square) of the table)
n number of plots which are the basis of the values compared.
The limits of reliability for the difference among means may be calculated in
accordance with the following formula:
D - ( ai - a 2 ) t t.sl/1 + !_
V n l n 2
where
HI and &2 m estimate of means of a^ and &2 treatments
t the value of Student's t at a given level of probability
S - test errors (residual MS of the table)
n 1 and n 2 - number of observations forming the basis of each of the estimates
of the means a^ and d2*
Random block design
Variance analysis of a random block design is not much more difficult than the
complete randomization test. There are two systematic sources of variation which can
be identified; these are blocks and treatments.
The mathematical model takes the following form:
1J
a
where
individual observation
- mean total
effect of the treatment 1 th
- effect of the block j th
- residual effect
(i -
(J
I)
A summary variance analysis table is shown below.
Source of variation
df
cs
F
Treatments
1-1
*i < x i.- x ..> 2
4
Blocks
3-1
I ^ (X tj - X.J*
Residual
(i-llj-l)
H<vv-v*.> 2
Total
u-1
55 <V X .. )2
In cases where, at a given level of probability, there are significant differences
among treatments, the real difference of D can be calculated and limits of reliability
established for the differences.
- 95 -
The test appears in the last column of the variance analysis table.
Latin square design
There are three identifiable sources of variation, which are lines, columns and
treatments, and it should be noted that the number of lines, columns and treatments is
equal. Through this design it is possible to isolate in the calculations variations
occuring both between lines and between columns, that is to say, in comparison with the
random block design, a further division has been made of the systematic site variations.
As may be seen from the variance analysis table below, the fairly strict design means
that the number of degrees of freedom for the variance of errors is reduced in compari-
son with the random block design.
The mathematical model takes the following form:
ijk
f Ak + fi
individual observation
mean total
effect of the line i th
effect of the column
effect of the treatment k th
residual effect
where
a - 1,
(k = 1,
I)
A summary of variance analysis table is shown below.
Source of variation
df
CS
F
Lines
(1-1)
X i !..-../
-
Columns
(i-i)
I 1 (X - X ) 2
J J
*
Treatments
ff.-D
I I (X - X ) 2
K. K.
4
Residual
a-Da-2)
9 2
Total
i 2 -i
A %K- ...'
* Since the values represented by i, j and k are aggregated,
all possible combinations will not be found.
The F test appears in the table; the real difference of D and the limits of
reliability are calculated as usual through standard error.
- 96 -
Incomplete block design
Incomplete block designs, particularly lattice designs, are relatively difficult
to analyze. Since they are used for fairly large and complicated tests, it is best to
use electronic data processing.
A brief description is given below of a general "intrablocfc" analysis, which can
also be used for different lattice designs.
The mathematical model takes the following form:
x ij M + a i + PJ + c ij where
tjj - individual observation
/Li - mean total
a - effect of the block i th
04 - effect of the treatment j th
fiij m residual effect
N - I.K. , where K - number of plots per block
A summary of variance analysis table is shown below.
Source of variation
df
CS
F
Blocks
l-l
K E (X. - X ) 2
i JL
Treatments fcorrected)
M
K Q* / X I
1
Residual
N-I-J+1
(difference)
1
Total
N-l
if tx ir x .. )2
The following comments may be made on the summary:
The corrected sum (CS) of treatment is calculated in accordance with the
following formula:
J *
A - the number of times that two treatments occur in the same block
Q J - T J -M n ij B i wh re
T. - the sum of all the observations included in treatment j
B. - the sum of all the observations in block i
n. - 1, if the treatment occurs in block i, a - 0.
The F test and the calculations of the real difference of D and the limits of
reliability are made as usual.
- 97 -
In lattice designs, Instead of the "intrablock" analysis described above, an
"interblocks" analysis can be made, also using a variation that can occur between the
blocks. The fl interblacks ff analysis Is described in detail by Cochran and Cox (1957);
Burley and Wood (1976) explain the method of analysis in connection with provenance
tests.
The lattice design analysis will be considerably more complicated if there are
missing values, and in such situations it may be necessary to make the analysis as in
random block designs, in this case omitting the division into Incomplete blocks.
Split plot design
Split plot designs are composed both of main plots and of sub-plots. The main
plots can be arranged in different designs, but in the example given below it is
assumed that the main plots are grouped in blocks and arranged in a random block
design.
The mathematical model takes the following form:
X-MV" 14 + flj + fl. 4- A-H + t/Jir "*" filk ** -MV where
A J XV ! i J A, J ^ IV 9 XIV J. J IV
X^jk individual observation
jti - mean total
a - effect of treatment i th (1 1, .. I)
0.j effect of block j th (J - l f .. J)
error (A) (k - 1, .. K)
effect of treatment k th
lk interaction between A and B
e ijk * residual effect (B error)
The variance analysis takes the following form:
Source of variation
df
CS
F
Treatments (A)
i-l
JK I (X. - X ) 2
x x
4
Blocks
J-l
XK (X , - X ) 2
J J
Error (A)
(i-l) (J-l)
o
KFT /Y V .. V X.Y \
4*4> IA. .A. A . T X J
1 J i J i J *
Treatments (B)
K-l
IJ Z (X - X ) 2
K. A
4
(A) x (B)
(I-l) (K-l)
J ?i (x .i fc - x .i.- x ..k +x ... )2
4
Residual
(Error (B))
,i
difference
Total
IJ(K-l)
fff (X iJk- x ... )2
- 98 -
As may be seen from the summary, two variances of errors are used to test the
effects linked to the main plots and those linked to the sub-plots, respectively.
The real difference of D and the limits of reliability are calculated in the
same way, using different standard errors for main plots and sub-plots respectively.
The split plot design is useful in those cases where it is desired to achieve,
for the effects linked to the sub-plots, greater precision than that desired for those
linked to the main plots.
CORRELATION ANALYSIS
The simple correlation coefiicient is a measure of the degree of linear associa
tion between two variables. This can often be useful in forest tree improvement, or
for investigation of the correlation among properties within the same individual, or
for the investigation of the correlation of a property among different and often
related individuals.
The simple correlation coefficient (r) is estimated as follows:
where SP indicates the sum of the products of the corresponding values of x and y, and
SC X and SC y indicate the sums of the squares of the deviations for the values of x and
y respectively.
The value of r can vary between -1 and +1. The values of -1 and +1 respectively
indicate complete negative or positive correlations, that is to say, all the points
included in the bidimensional diagram are found in a straight line with a downwards
(negative correlation) or upwards (positive correlation) skew.
The statistical significance of a coefficient of correlation at a given level of
probability can be determined as follows:
The null hypothesis: the coefficient of correlation has the hypothetical value
of Q 0.
The equation is used to indicate the size of the test.
If a t table is consulted and the size of the t test with (n-2) degrees of
freedom turns out to be significant at a given level of probability, the established
null hypothesis may be rejected.
REGRESSION ANALYSIS
The relation between the specific values of a variable and the mean of all the
Correlative values of another variable that depends on the first, is indicated as a
regression of the second variable over the first.
Regressions have the following general form:
- 99 -
^i * a + |3x^ + GI where
y * dependent variable
a coefficient
coefficient (coefficient of skew)
x Independent variable
e - residual effect
Coefficients a and f) are estimated as
SC
a * y -b x
The significance of linear regression can be tested through variance analysis,
as shown in the following table.
Source of variation
df
CS
F
Regression
1
b. SP^
4
Residual
1-2
(difference)
Total
1-1
1, (y - y.) 2
In many situations it will be necessary to use a more complicated model,
including several independent variables (multiple regression) , and in which may
also be found terms of squares or products or different variables.
COVAR1ANCE ANALYSIS
As in variance analysis, there are a series of different models for covariance
analysis. We will give only one example: a model of complete randomization tests, in
which, as well as the single identifiable systematic effect, an additional datum has
been registered in relation to each measurement.
In covariance analyses an adjustment is made of the values of plots through a
covariable. In most cases these are adjustments compensating for existing site dif-
ferences, but they can also be adjustments to compensate for the initial value of
the material, etc.
The mathematical model takes the following form:
1 j
where
- 100 -
,th
individual observation
mean total
effect of the treatment i tn (i - 1, .... I)
coefficient
co variable
residual effect
total number of plots in test
A summary covariance analysis table is shown below:
Source of variation
df
SC
F
Treatments
i-l
sc fc - sc ^ - sc
treat. tot. res.
4
x (covariance)
1
KcOT "?!^"y yil ~ y "" 2/
Residual
N-I-1
SC - --, % 2
res. LL (y - y ) - SC
<
Total
N-l
sc tot -!?<*- *..> 2
By including the covariable in the analysis, it is possible, as is shown in the
table, to calculate a corrected MS error and MS treatment, after which an F test can
be made in the usual way.
VARIABILITY COMPONENTS
If in an analysis there are random variables, the estimated MS can be disaggre-
gated in one or more variability components, each one relating to an identifiable source
of variation. The following table shows the distribution of expected MS in a random
block design.
- 101 -
Source of variation
df
Expected MS (E(MS))
Treatments
1-1
*e + J ' a treatment
Blocks
J-l
'e^ 1 ** block
Residual
(I-1XJ-1)
a e
Total
IJ-1
treatment the treatments component
o block - the block component
o ^ the residual component
The Individual variability components can be calculated from the MS values
estimated In the analysis.
The most important application of variability components in relation to forest
tree improvement is the estimation of genetic parameters within populations, also for
forecasting the genetic gains that could be achieved by selection.
The present article will not discuss in detail the different genetic parameters,
since this subject is covered In the article on quantitative genetics.
- 102 -
BIBLIOGRAPHY
Brown, A.G. & Matheson, A.C., 1977. Statistics: Interpretation of Results, Inter-
national course in Forest Tree Breeding, Canberra, Australia.
Bur ley, J. & Wood, P.J., 1976. A manual on Species and Provenance Research with
particular reference to the Tropics. CFI, Oxford, England.
Cochran, W.G. fl Cox, M.G., 1957. Experimental Designs, John Wiley & Sons, Inc. London.
Freese, F., 1970. Metodos Estadisticos Elementales para Technicos Forestales, Centro
Regional de Ayuda Tecnica, Mexico/Buenos Aires.
Kempthorne, 0., 1975. The designs and Analysis of Experiments. R.E. Krieger Publishing
Company, New York.
Scheffg, H., 1959. The Analysis and Variance, John Wiley & Sons, Inc., New York.
Searle, S.R. , 1971. Linear Models. John Wiley & Sons, Inc. New York.
Snedecor, G.W. & Cochran, W.G., 1967. Statistical Methods, The Iowa State University
Press, Ames, Iowa, USA.
Putting him on the spot with some difficult questions....
(Experimental plantation of Bombacopsis quinata)
- 103 -
SPECIES AND PROVENANCE TRIALS I/
R.L. Willan
CONTENTS
Page
Objectives * 4
The need for species and provenance trials 104
The types, sequence and time scale 104
Control of trials and their follow-up 105
Site assessment 105
Homoclinal comparisons and the choice of species to test 106
Practical limitations 106
Phasing and time scales of species trials 107
Species elimination phase . 107
Species testing phase 107
Species proving phase 107
Phasing and time scales of provenance trials 107
Range-wide provenance phase 107
Restricted provenance phase 108
Provenance proving phase 108
Plot size, shape and competition 108
Experimental design ... ... 109
Cultural treatment and protection 109
Priorities in field assessment 109
Recording systems 110
International cooperation. ... 110
Summary ... ... ... HI
References. ... HI
I/ ^Most of the following lecture has been extracted from "Manual sobre investigaciones
de especies y procedencias con referencia especial a los tropicos" (Burley and
Wood 1979).
- 104 -
OBJECTIVES
As described in the lecture on tree improvement in relation to national forest policy*
the objectives of afforestation and hence of species and provenance trials, should be stated
precisely in advance, in terms of closely defined materials or amenities to be produced.
The produce required may have a variety of possible uses, e.g. a saw log crop from which thin-
nings can provide poles and firewood, or it may have a very precise use, e.g. a high quality
veneer for export. Alternatively, what is desired may be an amenity e.g. decoration, shade,
shelter or soil improvement.
THE NEED FOR SPECIES AND PROVENANCE TRIALS
Trials are needed whenever adequate information is lacking, either on the requirements
of the species, or on the characteristics of the site, or both. In such cases embarking on
afforestation schemes without a carefully planned and executed experimental programme has
often led to costly failures.
The choice of species and provenances to use for afforestation involves the extra-
polation of information from elsewhere. Climatic and ecological matching of a new site and
the original habitat of a species is rarely enough since it cannot reveal the adaptability
of the species to new conditions or its ability to grow satisfactorily on a range of sites.
When information is lacking, the best way to acquire it is through trials of a number of
species in small plots on representative locations within the area of the proposed afforesta-
tion project. Provided the locations are carefully selected to sample the range of planting
sites and are properly looked after, extrapolation of performance from small plots to the
whole afforestation area should involve far less risk than imprecise comparisons, based on
inadequate data, between widely separated regions of the world.
The advisability of species and provenance trials is now generally accepted, but the
need for their careful planning and for high standards of maintenance and assessment has
often been less appreciated. The trials themselves can be wasteful and misleading if badly
planned or executed; and a proliferation of plots, if they are ill-sited, ill-tended and
ill-protected, is no substitute for a small, wisely planned programme which is tailored
to the staff and financial resources available. The objective is to derive the greatest
possible information from a given cost or, put the other way, to obtain the desired informa-
tion at the lowest possible cost*
For species with naturally wide geographical or ecological ranges, provenance testing
is essential. It is easy to be misled in the comparison of species for afforestation if the
total range of intra-specific variation is not known.
THE TYPES, SEQUENCE AND TIME SCALE
The ultimate phase is, of course, the complete afforestation project where the source
populations are reduced to one or two provenances of one or a few species and where the
annual planting area is reckoned in hundreds or thousands of hectares. It must be recognised
that there is no standard procedure or time schedule for passage through successive stages
of testing; nor is there always a need to use every stage. However, the following distinct
phases are commonly encountered.
The species elimination phase, the mass screening of a large number of possible
species in small plots for a short period (1/10-1/5 rotation) to determine survival and
promise of reasonable growth. The species testing phase is assigned for the critical test-
ing or comparison of a reduced number of promising species in larger plots for longer
periods (k-% rotation). The species proving phase is designed to confirm, under normal
plantation conditions, the superiority of a few probable species. Three similar stages
apply to provenance testing for species with a wide natural distribution, a range-wide
provenance sampling phase, a restricted provenance sampling phase and a provenance proving
phase. Since tnese phases usually apply to species considered likely or promising, plot
sice and time-scale may exceed those utilised for the comparable phases of species trials.
- 105 -
CONTROL OF TRIALS AND THEIR FOLLOW-UP
The detailed planning and conduct of species and provenance trials is the responsi-
bility of the silviculturist or research officer* However, the head of the forest service,
or the policy-maker to whom the silviculturist is responsible must be capable of evaluating
the research, and of deciding whether it will answer the right questions, with efficiency,
and within the resources available.
To facilitate control the research staff must be required to produce and follow a
project control plan for each species and provenance trial* This plan must be checked
and approved before the project is initiated. After any amendments are made the forest
director should signify approval in writing; this imposes on him the moral obligation
to do his best to ensure continued finance and staff for the project. If there is no
promise of continuity in this respect, there is little point in starting the programme*
Once a project is initiated, however, regular reports should be required (usually annually)*
A pilot planting project may frequently form an essential intermediate step between
species/provenance trials and large scale afforestation* It will enable the forester to
determine optimum cultural and managerial techniques and to make the vital decision on
whether or not to proceed with the complete afforestation programme.
It must be remembered that in some cases the results of species and provenance
trials are applied to the second, rather than the first, rotation* For example, many
countries are already planting large areas of Pinus caribaea var. hondurensis Barr. &
Golf., mainly derived from the Mountain Pine Ridge, Belize. Indeed some countries have
initiated selective breeding programmes based on this material. Nevertheless other
provenances may be better or contain some valuable genes, and these countries should carry
out comparative trials to identify such provenances for later planting. There is a danger
that foresters who have a readily accessible seed source of a "satisfactory 11 species or
provenance may feel it unnecessary to test others that could be potentially better.
SITE ASSESSMENT
It is important to realize that each trial does not measure just the performance of
a given species or provenance, but its performance on a particular site and with a particular
cultural treatment. Each of these factors interacts with the other two* The recognition
and mapping of the major site types in a potential afforestation area is therefore an
essential preliminary to an efficient programme of species and provenance trials* It
permits the eventual extrapolation of results from the trials to unplanted areas of the
same site type*
It is not satisfactory to locate trials by scattering them at random throughout
potential afforestation areas. Such areas are unlikely to have uniform environmental
characteristics. In these circumstances the random location of trials may fail to sample
a widely occurring site type.
The procedure of assessment should, therefore, be a process of progressive division
and sub-division to arrive at environmental units useful for the planning and interpretation
of trials.
Each unit should be a 'site 1 as defined by Coile (1952): -
"an area of land with characteristic combination of soil, topographic, climatic,
and biotic factors".
The ultimate definitions of these sites will depend on the degree of variability
encountered. However, a generalized procedure can be followed, as follows:
Classify each proposed planting area by:
(i) Latitude - to the nearest degree. (Latitude is related to day length).
(ii) Rainfall - according to its distribution throughout the year e.g. uniform, one
dry season, two dry seasons* If the annual rainfall shows wide variation with-
- 106 -
in any category a secondary division according to mean annual rainfall will
be necessary*
Draw the boundaries of proposed plantations on maps or aerial photomosaics, prefera-
bly using a scale of 1 : 50,000. On these maps, or a series of transparent overlays, mark
the following, if known, in this sequence.
iii) Geological boundaries, or the boundaries of generalized soil groups.
iv) The boundaries of conspicuous geomorphic features e.g. flood plains, river
basins, undulating hills, escarpments, dissected plateaus.
v) The boundaries of major topographic categories e.g. valley bottom, slope, hill
crest, plateau.
The ultimate units in the classification are, therefore topographic.
HOMOCLIMAL COMPARISONS AND THE CHOICE OF SPECIES TO TEST
The close matching of natural habitat and site for species introduction does not
eliminate the need for trials, since, however accurate the formulae used, the adaptability
and plasticity of a species cannot be assessed without testing. Moreover, the natural
distribution of a species may be due as much to the incidence of fire, ecological compet-
ition or man's activity, as to the measurable features of climate and soil. Many species
perform strikingly better in a new environment than they do in their natural habitat -
Pinus radiata and Eucalyptus sallgna are excellent examples of this.
On the other hand homoclimal information gained from the performance of a species
as an exotic can be of much greater value, and a review of information from other
countries can often reduce the number of possible species considerably and lead to the
inclusion of valuable exotic land races in provenance trials.
Most obvious non-starters - e.g. Douglas fir at low altitudes at the equator, can
be eliminated from the start, but in less certain cases it may be better to allow a
species to eliminate itself in trial, than to eliminate it on theoretical grounds, only
to have it reconsidered later when these grounds are forgotten. On the other hand, the
re-testing of species that theoretically should have done well, but did not, is often
worthwhile, because improved techniques or the build-up of mycorrhizal populations may
reverse earlier failures.
It is usually desirable to include a 'standard* well known species of provenance
in trials, in order to have a reference point against which to judge the performance of
the unknown populations.
PRACTICAL LIMITATIONS
The size of a programme of species and provenance trials depends upon many factors,
including staffing, finance and the availability of land, not only for the trials, but also
for subsequent planting on an operational scale* Security of tenure for the trial areas,
and public cooperation in the protection of the trials themselves, are essential. The
availability of suitably qualified staff will govern the kind of work that can be under-
taken, and training programmes may be a necessary part of the project. The availability
of transport may be critical in some countries.
The programme of operations must be worked out carefully in advance and costed as
far as possible. The total cost must cover not only the initial expenses but also the
essential maintenance of the plots throughout the duration of the experiments.
When staff or funds are limiting (and they are almost everywhere) the work needed
must be limited in advance, and in some circumstances an inadequately replicated series
of trials may be better than nothing. A precise estimate of species-site interaction,
on the other hand, costs money and this must be budgeted for.
- 107 -
The need to limit the research programme to a practical size makes it essential to
locate trial plots to ensure:
(a) Representative coverage of the main site types
(b) Easy accessibility for maintenance and assessment.
A larger number of small plots necessitates a long total perimeter, and hence more
costly protection. A high degree of replication and numerous species involves meticulous
labelling and supervision which may be difficult with relatively untrained field staff.
The propaganda value of a set of vigorous plots within sight of a road may be considerable
and, from the practical point of view, the simpler the statistical design and layout, the
better.
PHASING AND TIME SCALES OF SPECIES TRIALS
Species elimination phase
Object: To compare the performance of a large number of different species on one or a
number of sites, and to select a smaller number for more intensive trials.
Features: The individual species unit, or plot, is kept as small as possible.
The duration of such trials is commonly 0.1 to 0.2 x rotation age, and perhaps 20
to 40 species could be tested in the initial stages, though continued introductions of small
number of species are often made over a number of years.
Species testing phase
Object: The comparison of a restricted number of promising species, based on previous
experience, on sites within a broad climatic region.
Features: Properly designed statistical layouts are particularly important, and plots
must be of a size to enable reliable assessments to be made up to, at least,
the first thinning.
The duration of these trials may be about 0.5 x estimated rotation length.
Between 5 and 10 species is suggested at this stage.
Species proving phase
Object: To confirm, under normal plantation conditions, the results shown by a small
number of species that have shown themselves superior in earlier phases.
Features: Plots must be large enough to provide data on growth and yield for the full
rotation, surrounds must be large enough to eliminate or minimize edge effect.
In addition to 'normal 1 plantation methods, a range of other management tech-
niques may need to be tested, always in statistically valid designs. It is
also appropriate to investigate wood quality at this stage.
PHASIC AND TIME SCALES OF PROVENANCE TRIALS
The 'ideal 1 sequence of provenance trials follows very closely that outlined for
species above. They may be described as the Range-wide provenance phase, the Restricted
provenance phase and the Provenance proving phase.
Range-wide provenance phase
Object: To determine the extent and pattern of variation between provenances (populations)
of promising species with wide natural variation.
- 108 -
Features: Depending on the geographical distribution and variation of the species, 10-30
provenances are suggested at this stage. It often indicates groups of promising
provenances, and also areas from which large scale seed imports should be
avoided*
This phase is often run concurrently with species elimination or testing. Plot size
should be small but adequate for a duration of 0.25 to 0.5 x rotation age.
Restricted provenance phase
Object: To find sub-regions and ultimately provenances most suited to the sites under
test.
Features: The differences to be detected between provenances may be relatively slight, and
experimental design roust take account of this. Generally, 3-5 provenances may
be expected, with a duration in excess of 0.5 x estimated rotation, using plots
of the appropriate size. This phase is often run concurrently with species
testing and species proving phases. Local land races and other derived prove-
nances should be included where possible.
Provenance proving phase
At this stage one or two provenances only will have been selected for each species,
site and end use. The procedure is the same as that described for species.
PLOT SIZE, SHAPE AND COMPETITION
The size of plots, as indicated above, depends on the duration of the trial and the
expected growth rate of the trees*
As expense is also an important factor, it should be remembered that trials are
grown to provide information, and that this may be obtained as well from small plots as
from large. The single tree plot is of limited use except where a large number of species
or provenances is being screened for early survival, though it is cheap in space and cost,
and lends itself to a high degree of replication. Assuming an initial spacing distance
of 2 to 3 metres the following numbers of trees per plot are recommended*
Species elimination phase
Minimum plot size: 5 tree line plot; Maximum: 25 trees (5 x 5). No surrounds.
Species testing phase
Plot size 16-25 trees (4 x 4 or 5 x 5) with a 1 or 2 row surround.
Species proving phase
Because yield estimates are important, a central plot of 100 trees (10 x 10) plus a 2 row
surround may be considered a minimum.
Range wide provenance phase
25 tree plots, no surrounds.
Restricted provenance phase
25-49 tree plots, 1 or 2 row surround.
Provenance proving phase
100 tree plots (or more), 1 or 2 row surround.
Plots are usually square or rectangular, but may need to be elongated to fit
certain site configurations.
- 109 -
EXPERIMENTAL DESIGN
This is covered in a separate lecture. The most commonly used design is the Random-
ized Complete Block. It is simple, flexible and robust, but is less suitable if there are
very many species or provenances to be compared.
CULTURAL TREATMENT AND PROTECTION
'De luxe 1 treatment, that is measures to combat local hazards, such as intensive
weeding or the use of insecticides against leaf cutting ants or termites, should be used
in order to reduce as far as possible the chance of cultural methods masking the effects
of other site characteristics. Generally, therefore, competing vegetation should be reduced
to a minimum and the type of planting stock and method of planting should be the optimum
for the area concerned.
Clear prescriptions must be included in the experimental plan on such items as the
need for fencing or fertilization, time and method of planting, replacement of casualties,
weeding, protection against fire and pests, pruning and thinning. At planting it is
essential to ensure that plants are clearly labelled at all times and that planting is
organized to prevent the confounding of species/provenance plots with individual workers
whose planting skill may vary.
PRIORITIES IN FIELD ASSESSMENT
Assessment is time-consuming and expensive; therefore the characters to be assessed,
and the timing of assessments should be laid down in the control plan.
The initial calibrators of the field stage are assessment of height a few weeks
after planting (to allow the soil to settle) and a survival count.
The following are notes on the most important assessments.
Character
Frequency stage
Method
Health
continous
Note incidence of pests and diseases
Identify pests and pathogens.
Survival
1 yr. old. Subsequently
after climatic extremes,
etc.
100% count.
Sample count.
Mean height (h)
Annual up to about 7 m
height. Then every
2-5 years
Use measuring poles up to 7 m then
by optical instruments. Accurancy:
aim at 5%. 100% in initial stages,
or sample.
Dominant height (h dom )
(mean height of 100
trees of largest
diameter per ha)
Annual up to 7m height.
Then every 3-5 years
As above, but less time consuming.
Accurancy: as above.
Mean diameter (3)
at breast height
1.3 m from ground.
Annually after crop is
2-3 m high.
Diameter tapes most convenient, or
calipers. 100% sample essential
Accurancy: aim at 1-2 mm.
Basal area
Obtained from measurements of d.
Stem from Branch site
and angle
Start when trees are 7 m
tall. Then at 3-5 year
intervals.
Simple realistic systems best, using
scoring, 1-7 basis.
Bark thickness
Whenever d is measured
and on thinnings
Bark gauge, or bark removal on
thinnings; sample 5-101 of trees.
- 110 -
RECORDING SYSTEMS
A clear and accurate data recording system is essential for all experimental work.
Each item of data must be clearly identified and errors in recording and transcription
occur all too easily. Most commonly, data are recorded by hand on specially designed forms,
and these should be simple to use and provide a permanent record. The data should be
analysed as soon as possible, as the results may affect the future management of the trial.
INTERNATIONAL COOPERATION
The transfer of forest reproduction material, throughout the world, has reached un-
precedented heights during the past few decades; many countries have been trying to inten-
sify their forest production by introducing new species of provenances (Lacaze, 1978).
At the same time, there has been an increase in organized research aimed at supplying
silviculturists with objective information on the vegetative material to be selected.
An effort on this scale is inconceivable without active international cooperation.
Some progress has been achieved, but much remains to be done in this field.
During the past ten years numerous projects have been put into operation, very
often on the basis of international collaboration. In this connection special mention
should be made of the work done by various bodies such as the 1UFRO working groups on
provenance seed collection of North American species, FAO, the Commonwealth Forestry
Institute, the Danish FAO Forest Tree Seed Centre and the Centre Technique Forest ier
Tropical (France).
The number of species on which provenance trials are conducted is very high: just
as examples, we can list:
- For species concerning temperature zones: Abies grand! s Picea abies, Picea
sitchensis, Pinus contorta, Pseudotsuga menziesii, Larix europaea and Populus
trichocarpa.
- For species concerning Mediterranean zones: Pinus halepensis, Pinus brut la,
Eucalyptus camaldulensis, and Eucalyptus dalrympleana.
- For species concerning subtropical and tropical zones: Cedrela odorata, E. camal-
dulensis, E. deglupta, E. microtheca, E. tereticornis, E. urophylla, Pinus cari-
baea, P. kesiya, P. oocarpa, P. pseudostrobus, Tectona grand is, Terminalia ivoren-
sis and T. superb a".
In most cases the species studied are exotic to the country in which the experiments
are being conducted.
The many experimental plots (past and future) set up using material collected
within the framework of international efforts are or should be analysed in a cooperative
manner. This entails an effort of coordination at all stages:
- Establishment of experimental designs which are, if not identical, at least compa-
rable with each other in the various zones and countries.
- Development and utilization of identical procedures for measurement and obser-
vation. A particular effort should be devoted to the methodology of observations
on phenologicai stages, pest and disease damage, and measurements of wood quality.
- Establishment of coordinated timetables for measurements and observations.
- Utilization of modern means of data processing, at least at regional level.
In addition to the possibility of checking the results of each experiment against
those obtained in all the others, this method also has the advantage of making it possible
to estimate the effects of the interaction between genotype and environment and to iden-
tify "plastic" population, i.e., those adaptable to a large variety of ecological situa-
tions.
- Ill -
In addition, such coordinated action also has the great merit of enabling countries
or regions with limited research, infrastructures or equipment to benefit from the resour-
ces available and the experience acquired in other countries (Lacaze 1978).
Recent examples have demonstrated the effectiveness of the assistance that can be
provided through the implementation of programmes of this kind.
In countries where there are a considerable number of programmes for species and
provenance trials and conservation plantations, it is suggested that simple catalogues
be prepared on the work done and the results achieved, for exchange between specialists*
Finally, at world level, a publication such as "Forest Genetic Resources Informa-
tion 11 , edited by FAO, constitutes an interesting initiative for diffusing the most impres-
sive results of the main programmes.
SUMMARY
The objectives and methodology of trials have been well summarized by Eldridge
(1979). It is necessary to have:
(1) Objectives clearly stated;
(2) Seed collected by a reliable worker, preferably the investigator himself,
and well documented;
(3) Seedlings grown in the nursery under uniform conditions and with suitable
replication;
(4) A sound statistical design;
(5) Plot size determined by the expected variation in the material and by the
anticipated age of final assessment;
(6) Planting sites representative of future planting areas and as uniform as
possible;
(7) Great care in labelling and recording at all stages.
In addition the cooperation of countries in international trials allows a more
rapid and more efficient accumulation of knowledge at a reduced cost to each country.
REFERENCES
Burley, J. Wood, P.J. Manual sobre investigaciones de especies y procedencias con refe-
(1979) rencia especial a los tropicos. Trop. For. Paper No. 10, CF1 Oxford.
Eldridge, K.G. Provenances and provenance trials. In selected reference papers, Inter-
(1977) national Training Course in Forest Tree Breeding. Aust. Dev. Ass. Ag.
Canberra.
Lacanze, J.F. Progreso alcanzado en la seleccion de especies y de procedencias. Unasylva
(1978) Vol. 30 Nos. 119/120, FAO,Rome.
Coile, T.S. Soil and the growth of forests. Advances in Agronomy Vol. 4: 329-98.
(1952)
- 112 -
SEED STANDS
Mar cell no Quijada R.
Institute de Silvicultura
Universidad de Los Andes
MSrida, Venezuela
CONTENTS
Page
General remarks 1 ! 2
Selection of site 11 2
Assessment of area 113
Size and density of stand . 113
Genetic gains 114
Layout of area 114
Conformation of stand 114
Management and upkeep of stand 115
GENERAL REMARKS
Seed stands are areas selected in natural formations or plantations to ensure a
supply of seed of known geographic origin and parentage.
Seed stands constitute a reliable source of seed of certain genetic quality,
variable according to the quality of the formation, until such time as seed orchards prove
necessary and come into production* Since they constitute a local source, usually in or
near plantation areas in which a certain natural selection has already occurred, they
offer the advantage of providing seed that is genetically more reliable for the site than
seed brought from outside or from different environmental conditions.
Seed stands are a stage prior to the formation of seed orchards. The areas selected
for seed stands, unlike those chosen for seed orchards, were not originally intended for
this purpose: they acquire this function owing to certain characteristics of the formation
and the needs of the plantation programmes. The selection differential is usually lower
than in seed orchards and the rules concerning the trees present are less strict.
Genetic improvement through seed stands depends on the quality of the formation
and on the characteristics under consideration. The trees in the stand are not usually
subjected to progeny trials, so that their real genetic value is not known.
SELECTION OF SITE
The best natural or artificial stands of a given species are located where conditions
are similar to those of the area in which the seed is to be used. In selecting these
stands account is taken of the development of the formation (quality of the individuals)
and of the capacity of the wooded area available to produce enough seed to meet the minimum
requirements of the plantation programme or programmes.
In the case of plantations, particularly of exotic species, age is an important
factor in selection of the site, since the trees must be mature enough to guarantee the
production of viable seed in acceptable quantities within an acceptable period of tine,
- 113 -
but also have deep, wide, green crowns for maximum production of seed in the future. Often
the areas of physiologically mature or nearly mature trees available are very small; and
where the areas are sufficiently large, they have been planted only recently. Areas of
high mortality are usually discarded for seed stand purposes. For some exotic species
in zones with a difficult environment (such as high altitude or low temperatures, poor
soils or low rainfall), there is the problem that even when the species grows, repro-
duction is seriously affected, which is one constraint to the establishment of seed stands*
ASSESSMENT OF AREA
Once the area has been selected, the phenotypic quality of the individuals and the
initial average production are assessed morphologically and quantitatively. This will
enable us, in addition to laying down criteria for establishing the seed stand, to have a
basis for calculating the genetic advance to be expected in plantations grown from the
seed produced in that stand.
Morphological assessment is based on a few characteristics -3 to 5- considered to
be of importance for use of the plantation product. For timber and wood-based products
straightness of the trunk, forking and anomalies (particularly in conifers) are frequently
used, with different scales or categories. Some characteristics may be decisive with only
two criteria: rejection or acceptance. For example, trees with visible signs of attack
by certain insects or diseases are completely rejected, because of the ease with which
susceptibility can be inherited and the decisive importance of this character for the
success of future plantations. For the other characteristics, it will depend on the
quality of the formation how rigorous the selection is.
Quantitative assessment is made by measuring height and diameter and evaluating
the form and quality of the trunk.
SIZE AND DENSITY OF STAND
For a given quantity of seed required in a given programme, the size of the area
will depend on the productivity of the species and the density of the stand used.
The productivity of a species is given by the number of viable seeds per tree.
This is dependent upon the number of fruits per tree, the number of seeds per fruit and
the viability of the seeds. Prior sampling is essential to provide an estimate of
productivity.
The density of a stand depends on its age and/or development. In plantations,
between 100 and 300 trees per hectare are generally used, depending on the species and
on the environmental conditions. Competition between the crowns is always avoided in
order to ensure the maximum number of reproductive buds.
In tropical natural areas the possibilities of establishing seed stands are limi-
ted owing to the heterogeneity of the stands and the limited number of each species per
hectare, except in some specific types of forests.
Seed stands are characteristically monospecif ic, since the movement of pollen
between trees of a species is facilitated if the other species are eliminated. However,
in natural stands the possibility could be considered of combining two or three taxo-
nomically very different species so as to avoid problems connected with hybridization.
In this way more effective use can be made of an area devoted to producing seed.
For reasons of cost and productivity it is considered that an area of less than
5 hectares is not suitable for the commercial production of seed. Above this minimum,
the size will depend on the requirements of the programme, the original area covered by
the species, and the resources available for effective management of the area.
- 114 -
GENETIC GAINS
The main value of a seed stand is that it constitutes a local source of seed of
sufficient quality and quantity to minimize or eliminate dependence on external sources
that are quantitatively and qualitatively less reliable.
Any gain that may be achieved will depend on the quality and original conforma-
tion of the stand. The more regular the distribution of qualities among individuals in
the stand, the more opportunities there are for a selection that will lead to an impor-
tant genetic improvement, since there is more likelihood of all the desirable combinations
being achieved in one or few individuals. The more uniformly one or another type of
quality is distributed, the less advances are to be expected.
In first-generation plantations, particularly of exotic species and under some-
what limiting environmental conditions, distribution is often observed to be slanted
towards poor quality, with very few good individuals. In this case the greatest gain
to be hoped for is in survival, with a slight or regular improvement in development
quality. The improvement in survival is explained by the fact that the seed comes from
individuals that have survived well for a first-generation plantation, showing a certain
degree of cultural adaptability to the site, demonstrated by the fact that they have
achieved an acceptable stage of development under certain limitations.
Improvement in development is expected because, although the structure of the
stand and the need for a minimum number of trees per hectare may lead to the selection
of trees that fall outside the category of "good", the latter will be relatively more
frequent in the new (select) population and hence have a greater genetic impact.
LAYOUT OF AREA
A seed stand consists of a production area, or effective area, and a buffer area.
The production area is the area where the seed will be collected and which may
receive the cultural operations needed to promote production.
buffer area serves to limit contamination by pollen from uncontrolled external
sources, since seed stands are usually established in plantation zones. The buffer is
composed of the same species as the stand and therefore contributes pollen to the pro-
duction area. The same criterion regarding the effects of selection of individuals is
used as in the production area, but no cultural operations are undertaken.
buffer area will be between 100 and 500 m wide, depending mainly on environ-
mental factors that favour dispersal of the pollen.
CONFORMATION OF STAND
Once the area has been selected and assessed, the guidelines for intervention
and conformation of the stand are established. Individuals should as far as possible be
distributed regularly over the whole area, with a minimum space between them to encourage
flowering and pollen dispersal. The minimum space is that necessary to avoid precocious
competition and may be as much as 1.5 times the initial spacing in young stands. In
any case, spacing will depend on conformation of the neighbouring trees and the density
adopted.
In plantations work can be done systematically by rows, thus encouraging regulari-
ty of the resulting stand. Since the aim of selective thinning is to leave only one quar
ter to one sixth of the original crop per hectare, it is often done in stages (2 to 4
years), so that any necessary corrections can be made to the initial operation.
- 115 -
of the area the first year and the remaining three quarters in the other two years.
The time taken to establish a stand usually depends on the area to be covered and
the constitution of the initial formation (density and size of trees).
It is essential that the project leader mark the trees to be operated on so that
at the same time as he/she trains the auxiliary staff who usually complete the work, the
necessary correction can be made. Technical supervision is necessary in any case in order
to guarantee good results. Marking should always be done well in advance of tree clearing,
and this should be taken into account in planning activities.
Felled trees must be removed from the stand in order to eliminate obstacles and
possible sources of attacks by pathogens.
The stumps should usually be fumigated to prevent them becoming foci for insects
and diseases.
MANAGEMENT AND UPKEEP OF STAND
The main management activity in a seed stand is thinning. Protection against fire,
insects and diseases is also of great importance. The maintenance of firebreaks and
adequate access roads permits effective mobilization in the event of fire. The elimination
of low vegetation and control of stumps also helps to avoid possible centres of infection.
Biological and genetic research work, such as studies on the use of fertilizers
or other means of encouraging flowering and fruiting, can also be conducted in seed stands.
- 116 -
SELECTION AND MANAGEMENT OF SEED STANDS
WITH SPECIAL REFERENCE TO CONIFERS
W.H. Barrett
Fiplasto S.A., Buonos Aires, ARGENTINA
CONTENTS
?age
Introduction 1 16
Background and present situation in Latin America , 117
Location and selection of seed stands 118
Selection criteria 118
Characteristics of the stand 119
Age 119
Area 119
Isolation 119
Accessibility " ... 119
Stand management 119
Seed production 120
Genetic gains 120
Bibliography 120
INTRODUCTION
The use of seed stands is an interim measure in most forest improvement programmes,
which^enables improved seed to be obtained in a short period of time. It is considered an
interim measure because the intensity of selection from stands can never be the same as
that from seed orchards. There will therefore be less improvement, or rather, fewer genetic
gains .
In regions where large-scale afforestation programmes are undertaken, as is usual
in tropical or sub-tropical areas, there is a great demand for seeds. The lack of sufficient
quantities of seeds is often a bottleneck in plantation planning.
The source from which the seed is extracted, (its provenance) is important to the
future of the plantation. This source will influence not only the volume to be obtained,
but also adaptation to the plantation site, plant health, type and rate of growth, wood
quality, etc.
Seed stands can be selected both in native forests within the range of natural
dispersion of the species, and in plantations within and outside the area of origin.
This paper is based on the lecture given by H. Keiding at the FAO/DANIDA Training Course
on Forest Seed Collection and Handling, Thailand, 1975, with additional information from
H. Barner, from the FAO/DANIDA Training Course on Forest Tree Improvement, Kenya 1973.
- 117 -
Various names have been given to seed stands, particularly in English-speaking
countries, such as "seed source", "selected stands", "seed stands", "seed production areas"
which are discussed in detail by Barner (1973). A distinction must be made between stands
selected for the general quality of the trees and stands in which intensive thinning has
been carried out, eliminating inferior individuals.
The following definitions are adopted in accordance with the OECD (Organisation for
Economic Cooperation and Development) scheme:
Stand; "A population of trees possessing sufficient uniformity in composition,
constitution and arrangement to be distinguished from adjacent population".
Seed stand/Seed Production; "A plus stand that is generally upgraded and opened by
removal of undesirable trees and then cultured for early and abundant seed production".
The objectives of creating seed production areas of seed stands are, according to
Matthews (1964):
1) To proudce seed of improved quality through selection and elimination of inferior
trees, favouring those that are vigorous, straight-stemmed, healthy and producing
wood of good quality.
2) To concentrate seed collection in a few specially treated parts of the forest,
thus making seed collection easier to organize and control.
3) To improve the germinative energy and germinative capacity of the seed collected.
BACKGROUND AND PRESENT SITUATION IN LATIN AMERICA
The selection and establishment of seed stands is the quickest method of obtaining
any improvement in the quantity and quality of seed harvested. It would appear that integral
and long-term improvement procedures include the selection of seed stands to obtain improved
seeds, until seeds can be obtained from more carefully selected trees in seed orchards.
At present, seed stands are sometimes still the only way for some countries to obtain
seed for their plantation programmes. This was the practice most frequently used in the past.
Stands are usually selected from mature forests, and subsequently treated to trans-
form them into seed producing areas.
Forest gains or improvements, and the efforts needed to achieve them, will depend
on the nature of the basic material one has to work with.
Following Bamer (1973) and Jones and Hurley (1973), the basic u>aterial has been
classified in accordance with its origin and location, as follows:
1) Mainly located within the range of the species, whether in natural stands or
plantations.
This group, particularly the conifers, includes such countries or regions as the
United States, Canada, Mexico and countries of Central and Southern Europe*
2) Mainly exotic species, introduced into the region, where they have been cultivated
fairly extensively.
This group can be found in almost all the countries whose forest production is based
en extensive cultivation of exotic species.
Conifers, particularly, nan be found in countries such as New Zealand, Australia,
Chile and Spain, where Pinus radiata has been extensively cultivated; in South Africa,
Zimbabwe, Malawi and Kenya (Pinus patula); in Argentina, Brazil and Australia (pines froip
the southeastern United States); Great Britain and central western Europe (Pseud ot suga menzlesii
and other species of pines), to give only a few examples.
- 118 -
3) Mainly in experimental plots of uncommon or recently introduced species.
This group includes most tropical and sub-tropical countries, and also many species
now being tested in developing regions. It applies particularly to species trials and
prevenances within species which have performed outstandingly in trials as compared with
the material widely used throughout the region.
As regards the use of seed stands of native species in tropical regions, the limited
use of this system can be explained by the fact that native species are usually replaced by
fast-growing species which are almost always exotic. Even when some native species can be
used, the mixed nature of the forest makes it very difficult or practically impossible to
transform a forest into a seed stand.
Tropical areas in Latin America are no exception to this rule for native species.
Species are scattered throughout mixed forests, where a large number of species, mainly
hardwoods, grow side by side. Native conifers occur in climates ranging from sub-tropical
to cold temperate. Apart from some exceptions, such as certain pine species in Mexico and
Araucarias in Argentina, Brazil and Chile, most conifer species are not cultivated and are
Araucaria angustifolia. In the Argentine province of mi si ones alone, scape 30 000 hectares
have been planted. So far, seed has been collected from native stands - although seed
stands are now being developed.
Group 2, as mentioned above, includes Chile, Brazil and Argentina. The questionnaire
to be completed by course participants will probably provide further inf anna t ion of species
cultivated in other American countries.
The remaining species used on a large scale in many American countries are still
included in group 3, since there are no forest plantations old enough for selection of seed
stands. The oldest and most successful experimental plantations are used for this purpose.
LOCATION AND SEUSCTION OF SEED STANDS
Here again, a distinction must be made between native species in their native area
and exotic species widely cultivated and recently introduced.
In the first case, selection will have to be based fundamentally on existing infor-
mation from the various provenance trials which determine suitability for climates and
soils. Chce the region of area has been determined, a stand will be selected for optimum
wood quality and quantity, and quality and regularity in seed production.
As regards exotic species, a distinction must be made between those widely cultivated,
such as Pinus radiata, . caribaea, . elliottii and IP. patula in the southern hemisphere,
and those not widely cultivated or only recently introduced. The methodology indicated
below refers to the first group, where selection of stands will have to be stretched as far
as possible using all available information on the value of the seed source, reproductive
quality of the stand under consideration, results obtained in previous plantations, etc. As
regards the second case (species not widely cultivated), progress will have to take its own
course, little can be done about it.
The total area needed for seed stands for any given region depends on the annual
demand for seed, productivity of stands, by tree or by area, and quality of production.
Selection criteria
In selecting the seed stand, its uniformity and volume production must be borne in
mind, and the trees must have well-shaped trunks, good growth habit, quality of wood and
good health.
In European countries (for example, Sweden) stands are graded into "plus", "normal"
and "minus", there being a tendency to subdivide the plus stands into different categories
depending cm their degree of superiority. In other countries (for example, Great Britain),
the stands is classified on the basis of individual trees. Samples are taken systematically,
within which all the trees are measured and classified. The total of points in the simple
determines the total points of the stand, according to which it is declared plus, almost
plus, normal or minus.
- 119 -
Although they appear different, selection criteria in both cases are similar with
respect to the characteristics chosen to classify the trees. Selection criteria are also
similar to those of individual phenotypic selection, but they differ fundamentally in the
intensity of selection.
Characteristics of the stand
Age
Age varies with species and region. A stand should be old enough to have proved the
value of its seed in areas where it was used before. It should be mature enough for good
flowering and fruiting. It should also be old enough to permit correct evaluation of stand
characteristics. Barner (1973), however, maintains that, due to possible difficulties in
evaluating some characteristics in older stands, it is desirable to start evaluation at an
early age, if the stand promises to be valuable, and keep this information on file and up
to date until needed. The seed stand, although it should be mature, should not be very old,
so that seed extraction can continue for a reasonable number of years.
Area
It is not possible to establish the area of a seed stand in advance. In general the
criteria in the minimum area in which it is economically worthwhile to collect seed. In
western Europe, 2 to 5 hectares are considered the minimum area for seed stands. The stand
should also contain enough individual trees of the desired species to allow for reasonable
cross pollination. This is particularly important in mixed forests.
Isolation
The isolation factor should be borne in mind, to ayoid contamination by inferior
stands. Wind -pollinated conifer species are more difficult to isolate than broad leaved
species, which are usually insect-pollinated. Isolation can be achieved by choosing stands
which are between 300 and 1 000 metres away from any possible source of contamination. In
some cases, plus stands can be found sufficiently far away from contaminating minus stands
for the forest itself to be used as the buffer. In this case, only the area of the plus
stand farthest from the contaminating source is harvested, using part pf the same stand as
a buffer zone.
Accessibility
Normally, easily accessible stands should be selected, to reduce the cost of manage-
ment, inspection, upkeep and seed collection. When genetically adequate species are found
off. the beaten track, it is preferable to collect their seed and sow it in places where
it is easier to reach.
STAND MANAGEMENT
The treatments normally given to a stand within an area to be used as a seed source,
and to convert it into a seed stand, are as follows:
1) Removal of minus trees to improve the genetic quality of the seed.
2) Thinning, to allow adequate space for flowering, fruiting and seed collection.
3) Clearance of the undergrowth to facilitate seed inspection and collection.
4) Demarcation of the stand, particularly when there are contamination problems.
5) Treatments to increase production, such as pruning or application of fertilizers.
6) Other treatments to protect fruiting, such as application of fungicides and
insecticides.
To improve management of seed stands, it is desirable to keep a record covering all
activities and treatments applied, pheno logical information, and information on
harvesting.
- 120 -
SEED PRODUCTION
Relatively little is known about seed production from a seed stand , as compared with
a stand that has not been treated. It is known that when a stand is thinned and the canopy
opened, an increase in seed production is immediately obtained. However, the duration of
this effect is not known. It is probable that there is an increase in production per tree,
but it is not known if there is greater production per area. An improvement in seed quality
is certainly achieved by eliminating minus trees and thinning to get more light to the
crowns of selected trees.
Nor is there any concrete information on increased production as a result of
fertilization, since almost all research has been done on seed orchards. In Japan, studies
made en Pinus densiflora and Larix leptolepis showed that the application of fertilizers
improved seed quality, and thinning produced an increase n seed quantity. The best result
was dbtained with light thinning; this was not only better than the control treatment (with-
out thinning), but also better than heavy thinning (Asakawa et al, 1969).
GENETIC GAINS
To date, there has been an increase in the quantity and physiological quality of the
seed produced in seed stands.
In species with wide variations in breed or ecotype, usually species with a large
range, maximum gent ic gains of the seed stand depend on the correct choice of the basic
population. The original source of seed should be known through previous studies of prove-
nance trials or as a result of intensive and extensive silviculture. For this, both the area
of origin and plantations established in the region with material from the same source can
be used. For example, in Argentina in provenance trials it was found that populations of
Pinus taeda from central Florida with 7-year rotation grew 200% in volume as compared with
populations from the foothills of Georgia. Something similar happened with the same species
in Queensland. As the basic population used for seed stands is from different sites in
Georgia, the production in volume can easily be improved by using seeds from Florida stands.
As regards characters within the population such as shape, width of crown, density
of the wood, etc., this will depend directly on the her it ability of the character and the
intensity of selection. This being a mass selection, with relatively low selection intensity,
genetic gains will depend on heritability.
Okie example from Shelbourne (1969) for diameter low (heritability) and bole straight-
ness (high heritability) in Pinus Radiata at a selection intensity of 1 in 10, expressed in
percentage of the average population before thinning, is 25% for bole straightness and 5.6%
for diameter. At a selection intensity of 1 in 20, the gains anticipated are 29.2% and
6.7% respectively.
In Latin American countries, where many factors relating to species, provenance, site
and their interactions, have not yet been explained, much greater gains may be hoped for in
the search for species and provenance better suited to each region. For this reason a tech-
nique recomnended is that of establishing simultaneously with species and provenance trials,
large-scale plantations of each of the biological units tested. This material could be used
in the future as a source of seeds in the short term, and selections could be made from it
for long-term improvement programme.
BIBLIOGRAPHY
Asakawa, S. & Keiji Fujita, 1969. Studies on the management of experimental seed stands pf
Pinus densiflora and Larix lep to lapis and the results obtained for three years (1962-
1964). Bull. Gov. For. Exp. Sta. No. 184, Tokyo, Jap.
ftarner, H. 1973. Classification of sources for procurement of forest reproductive material.
FAO/DANIDA Training Course on Forest Tree Improvement, Kenya, pp. 119-121.
E lias on, E.J. 1969. Development of seed production areas-an interim selection* FAOHEUFRO.
Washington. 2: 1367-1371.
- 121 -
Jones, N, & J. Bur ley, 1973. Seed certification provenance nomenclature and genetfc history
in forestry. Silvae Genetica 22 (3); 53-58.
Keiding, H., 1975. Seed stands. FAO/DANIDA Training course on forest seed collection and
handling - 2: 192-211 - Thailand.
Matthews, J.D., 1964. Seed production and seed certification. Unasylva 18 (2-3): 104-108.
Mittak, W.L. , 1979. Cursillo para el mane jo de rodales seleccionados para la produccipn de
semi lias. Institute Nacional Forestal BANSEFCR Proyecto GUA/78/005. Guateipala.
Shelbourne, C.J.A., 1969. Predicted genetic improvement from different breeding methods.
Second World Consultation on Genetic Improvement. FAQ-IU7RO Washington 2: 1023-1029.
- 122 -
SELECTION AND MANAGEMENT OF SEED STANDS; HARDWOODS
C. Palmberg
Forest Resources Division
Forestry Department, FAO
CONTENTS
Page
Introduction 122
Stand selection 122
Age 123
Area 123
Isolation 123
Stand Management * 123
Bibliography 123
INTRODUCTION
The purpose of a seed stand is to meet immediate needs of relatively large geneti-
cally somewhat improved quantities of seed by setting aside stands of not lower than
average quality and managing them with the primary objective of seed production.
General principles of selection and management of seed stands, as outlined by
Barrett in the previous lecture, are valid for hardwoods as well as for conifers. Points
on which methodology between hardwoods and conifers differ or which may require particular
attention when dealing with hardwoods, are outlined below. References to publications
dealing with the establishment and management of hardwood seed stands are listed at the
end of this note.
STAND SELECTION
Hardwoods are often dioecious. In selection of stands, the availability of both male
and female trees must be ascertained and when thinning, care must be taken to leave
a balanced sample of male and female trees;
In species with pronounced root-succering ability large areas of trees which are
genetically identical sometimes occur. For these species the seed stand approach may
not be suitable;
When dealing with plantation species it is very important that origin of the stand is
known. Exotic plantations may be based on seed from relatively few trees and although
the stand itself may be of superior growth and form, the genetic quality of the seed
collected from it is likely to be poor due to inbreeding effects. Sometimes, e.g. in
the case of eucalypts, the stand may be of hybrid origin, and seed collected from F,
hybrids and from subsequent generations will produce unacceptably variable stands
because of segregation of the genes;
- Selection criteria such as health of trees, yield and form are universal. However,
additional, species-specific selection criteria will often be included for hardwood
species; examples of such criteria are selection against genetically controlled
defects characteristic of some species, such as the tendency to produce epicormic
shoots having bud traces extending to the pith as in liquidarabar styracif lua and
against the formation of a large number of kino veins in Eucalyptus regnans; se lect ion
for extremely low wood density in Ochroma lag opus selection for specific wood
characteristics such as attractive colouring or patterns of wood in some high-
quality cabinet timbers; selection for high coppicing ability especially in some
species grown on short rotations.
- 123 -
AGE
The stands must be young enough to be able to respond to the first thinning by the
development of crowns capable of producing large seed crops, but old enough to give some
indications of the characteristics to be selected for.
AREA
The distances between individual trees of the same species are often large in forests
with a heterogeneous species composition. Provided these species are outbreeding and the
seed stand appearance therefore per s.e is valid, the areas needed are much larger than if
dealing with forests consisting of one or only a few species. If the density of the species
is very low, the seed stand approach cannot be used.
ISOLATION
Knowledge of the breeding system of a species is fundamental. If a species is in-
breeding, no isolation is naturally needed. However, forest tree species are largely out-
breeding, possessing various mechanisms to prevent selfing.
Effective pollen flight distances and the distance pollen will be transported by
insects are not well documented for many hardwood species. The isolation of hardwood seed
stands should be kept similar to that for conifers (- 300 m) until further studies have been
undertaken.
STAND MANAGEMENT
Seed stands should undergo a series of thinnings to increase quantity and improve
quality of the seed produced. The first thinning should be made before serious crown
competition sets in; the main purpose of this thinning is to allow the development of wide,
deep and dense crowns, i.e. to increase potential quantity of seed at the same time removing
inferior phenotypes. During the development of the stand, further variation will manifest
itself; as phenotypic selection will generally be more reliable in mature trees, the im-
provement achieved through thinnings will gradually shift its main emphasis from quantity
to quality.
Knowledge of the biology of flowering and seed production is essential for the mana-
gement of seed stands, and this knowledge is often scarce or lacking in hardwoods. It is
important to be able to forecast the exact time of the year that flowering will take place
so that thinnings can be timed to precede flowering any one year; incorrect timing may
cause a delay in the benefits of the thinning equal to the period between successive seed
crops. It is also essential to know the time required between pollination and the matura-
tion of seed so that the relative degree of improvement, which is directly related to selec-
tive thinnings, can be forecast.
BIBLIOGRAPHY
Melchior, G.H., 1977. Programa preliminar de un esayo de precedencia de Cordia alliodora,
Cupressus lusitanica y otras especies nativas y ex6ticas. Proyecto Investigaciones y
Desarrollo Forestales COL/74/005. PIF no. 7. Bogota, Colombia.
Melchior, G.H. & Venegas Tovar, L., 1978. Propuesta para asegurar el suministro de semillas
de Eucalyptus globulus en calidad comercial y geneticamente majoradas. Proyecto Investiga-
ciones y Desarrollo Forestales COL/74/005. PIF no. 14. Bogota, Colombia*.
van Dijk, K. , Venegas Tovar, L. & Melchior, G.H., 1978. El suministro de semillas como
base de reforestaciones en Colombia. Proyecto Investigaciones y Desarrollo Forestales
COL/74/005. PIF no. 13, Bogota, Colombia.
- 124 -
SELECTION OF FOREST TREES
N. Quijada R.
Instituto de Silviculture
Universidad de Los Andes
Mfirida, Venezuela
CONTENTS
Page
Genotype and environment 124
Factors to be considered in the selection of individual trees 125
Type and number of characters 125
Intensity of selection 125
Propagation of plus trees 126
Selection procedure 126
Method of improvement through selection . 126
Sources of trees for selection 126
Selection criteria 127
Selection terminology . 127
Bibliography 127
Annex 1; Example of selection criteria: selection criteria for seed stands ... 128
Annex 2; Example of selection criteria: selection criteria for seed orchards .. 129
GENOTYPE AND ENVIRONMENT
In practice, the commonest improvement methods are based on the selection of indi-
viduals meeting certain minimum requirements for specific characteristics which are impor-
tant for some specific purpose.
The external appearance, or phenotype, of an individual is the first guide for the
breeder. Although the phenotype is based on 2 components: genotype and environment, which
may or may not be of equal importance in the resulting appearance, the breeder bases him/
herself initially on the probability that a good phenotype has a sufficient genetic basis
to produce a reasonably favourable reaction in different environments.
Many characteristics are known to be genetically controlled, either partly or almost
entirely, so that external appearance reflects inherent potential to a large extent. Other
characters vary more in this reaction to environmental conditions, so that a specific
phenotype reflects only one of the many possible reactions.
The existence of different forms or alleles of a gene, and the occurrence of poly-
genesis, pleiotropy epistasis, etc. result in great complexity in the expression of a
character or various combinations of characters. The breeder selects from the whole range
of expressions, or variations, the most favourable types or combinations. The selection
practised by the breeder is controlled or truncated, in that he/she selects towards an
extreme of the original population and tries to shift the mean of the future populations
in that direction.
-125 -
The importance of the genetic component of a character can only be determined by
field trials, either through vegetative propagation of the individuals selected in
different environments, or through their progeny (progeny trials). Meanwhile, external
appearance will continue to be the basis for work.
The stricter the selection criteria, the more guarantee there will be of good genetic
gain. However, this depends on the species, the product and the quality of the population
available for selection.
An initial survey of the areas of distribution (plantations or natural forests) helps
to establish a criterion for the requirements of the various characteristics. This is
particularly important for selection programmmes with first-generation exotic or native
species affected by dysgenic selection over much of their range.
FACTORS TO BE CONSIDERED IN THE SELECTION OF INDIVIDUAL TREES
The success of selection based on phenotype expressed as genetic advance or gain,
depends on various factors, in particular the type and number of characters in the
selection, the intensity of selection and the method of propagation.
Type and number of characters
The type of character (high or low heritability) has a decisive influence on the
advance that can be achieved. Highly inheritable characters are easier to manipulate and
more predictable in their responses. Such characters are straightness of the trunk,
forking, and resistance to diseases. Characters with low heritability are less predictable,
since they require greater environmental control. They include important economic character-
istics such as physical and chemical properties of the wood.
The number of characters also affects the response obtained. It has been shown that
the greater the number of characters involved the more difficult it is to obtain advances
of specific individual characters. This is due to two factors: first, different characters
have different inheritance patterns that require different selection intensities. If the
number of individuals possessing a given character is increased, they will be increased
danger of introducing undesirable phenotypes of this character. Secondly, different
characters may be inversely correlated, so that greater strictness in one would have a
negative effect on the expression of the other.
It is therefore necessary to concentrate on a few characters at a time. First, one
chooses the most manipulable and the most important selection factors, such as straight
bole, forking, vigour, etc., and second, one considers the properties of the wood (5).
Intensity of selection
Trees are selected with two considerations in mind, among others: that the trees
selected have as little family affinity as possible, in order to avoid problems connected
with consanguinity; and that it be possible to obtain a minimum number of individuals, in
accordance with the purpose of the selection. As regards the first, in natural stands
distance is a reliable indicator. The closer the trees, the greater the probability of
consanguinity, simply because of how pollen moves. Widely spaced trees, either in the
same area or, better still, in different areas, show fewer family affinities. In plan-
tations we expect at least the trees planted in the same year to be more related to each
other. From one year to another, and one site to another, the specific source of seed
may vary, particularly when very large lots of seed are required. In this instance an
effort will be made to reduce the total number of selections on any one area of the
plantation.
On the other hand, the number of selections influences the range or genetic base.
A limited number of selections will create a very narrow base, which could rapidly lead to
problems, including problems of consanguinity. In seed orchards 20 trees are considered
an absolute Minimum for maintaining a sufficiently broad genetic basis to achieve important
advances, and to ensure that the seed produced can adapt to the natural variability of
plantation sites. This is based on the assumption of regularity in time and quantity of
- 126 -
flowering and fruiting of all trees. Since this does not always happen it is preferable to
use a greater number. In seed areas the number of trees left will depend on seed require-
ments, the development of the plants and the area available, but it is usually between 150
and 250 per hectare.
Strictness with regard to intensity of selection diminishes as parameters of varia-
bility, such as heritability, become available. Intensities of 1 tree per 700 hectares and
1 tree per 1.2 hectares have been used for Pinus radiata in New Zealand and Australia
respectively; 1 tree per 8 000 for Cupressus lusitanica in Colombia and 1 per 750 (1 per
0.65 hectare) for Pinus caribaea in Cachipo, Venezuela (4, 5, 8).
In any case the final decision will depend on the variability of the species and on
the immediate needs for seed in terms of quantity and quality.
The intensity of selection is determined in various ways. One is by means of a
selection differential expressing the difference between the mean of the original popula-
tion and the mean of the selected trees.
The mean of the original population can be estimated by sampling the population.
Since this may be expensive and time-consuming, the means of the 4 or 5 best trees near
the plus tree have been used as a population reference model.
With this value, and with the estimated heritability value (h ) for a particular
character, we can calculate the genetic gain (R) expected. (See the paper on quantitative
genetics for more detailed information).
Propagation of plus trees
In the case of trees selected for inclusion in seed orchards, there are two
alternative methods of propagation: sexual and asexual: which determine the type or orchard
to be formed (seedling and clonal respectively).
SELECTION PROCESS
Method of improvement through selection
The two most common methods of selection are stand selection and selection by
family (1, 10).
In stand selection, individuals are selected on the basis of phenotype and then
allowed to cross-fertilize freely. This is the usual practice in seed stands and seed
orchards, where the seed is mixed without concern for family relations. To test the
effectiveness of stand selection, a progeny trial may be conducted with groups of both
plus and normal trees from the original populations. This procedure has proven effective
for highly inheritable characters.
Selection by family makes it possible to control parental relationships in the
resulting progeny, thus facilitating continuous evaluation of the plus trees. The most
common family relationship is sibs or half sibs. Different procedures and cycles of
assessment, grouped under the generic term of recurrent selection, permit the elimination
of trees originally selected and the incorporation of new selections. The practice is
most common in seed orchards for production and control of development (with progeny
trials).
Sources of trees for selection
The source of the trees to be selected is determined by the use to be made of the
selections. In the case of seed stands, selection is made within a specific area, normally
a plantation of known origin, but sometimes a natural stand which has good overall
characteristics of vegetative and reproductive development and is large enough to guarantee
sufficient seed output. It must also be located in the zone where the seed will later be
used. For seed orchards, the total area of natural and artificial stands within a climatic
region is taken as the reference. In order to have a good initial basis for work, initial
election can be in the best stands of the best provenances.
- 127 -
Selection criteria
There are two practical methods for assessing trees: individual evaluation and
comparative evaluation. The first method consists of evaluating each tree on its own
merits, according to scales of values for the classes in the characteristics. The different
classes in each individual characteristic will be given by technical criteria, in accordance
with the phenotypic variants discernible in each one. This alone establishes a certain
level of sublectivity. This method of evaluation is appropriate in samples made in for-
mations that are to be used for seed stands, and in which it is desired to assess the
formation in order to intake decisions on the intensity of and criteria for interventions.
It is also used to evaluate the development of plants in progeny trials. In this system a
level is established beneath which any tree is automatically eliminated, regardless of the
values in other characteristics.
The second method uses scales of values resulting from the superiority of the
candidate tree over comparable trees in the vicinity. Usually a number of additional points
are assigned for so much percent or absolute value of superiority, the magnitudes depending
on the weight it is desired to give to each characteristic.
This works fairly well for characters quantifiable by specif ic units (height, diameter,
volume), but it is also used for more qualitative characteristics where somewhat subjective
classes are taken as a basis.
Some characteristics are determinant, regardless of the selection criterion. An
example is resistance to pests or diseases: usually any trace of attack automatically
eliminates the tree that is being assessed.
Terminology
A tree which appears phenotypically desirable and which is to be the subject of
evaluation is called a candidate or preselected tree. Once its characteristics have been
assessed and it has been accepted for further use, it becomes a select or plus tree.
One its superior genetic characteristics have been ascertained, it is known as an
elite tree. Several cycles of progeny trials are often required before a tree attains this
status.
BIBLIOGRAPHY
1. Allard, R.W., 1964. Principles of Plant Breeding. John Wiley & Sons, Inc. New York.
485 pp.
2. Brown, C.L. & R.E. Goddard, 1961. Silvical considerations in the selection of plus
phenotype. J. For 59: 420-426,
3. Falconer, D.S., 1960. Introduction to Quantitative Genetics. The Ronald Press
Company, New York, 365 pp.
4. Gutierrez, V., M. and W.E. Ladrach, 1978. Iniciaci6n de un Programa de Mejoramiento
Gene*tico de Cupressus lusitanica y Pinus patula en Colombia. Boletfn IFLAIC,
No. 53: 3*19.
5. Keiding, H., 1974. Selection of Individual Trees. In FAO/DANIDA Training Course on
Forest Tree Improvement, Kenya, pp. 165-175.
6. Ledig, F.T., 1974. An analysis of methods for the selection of trees from wild stands.
For Sci 20: 2-16.
7. Morgenstern, E.K. 1975. Review of the principles of plus tree selection. In Plus Tree
Selection: Review and Outlook. Publication 1347, Canadian Forestry Service,
Ottawa, Canada, pp. 1-27,
- 128 -
8. Smith, N., 1976. Selecci6n de irboles en Cachipo para estabiecer un huerto semillero
^ e P* nus caribaea More let. II Seminar io Nacional de Plantaciones Fores tales
S.V.I.F. MSrida, Venezuela. 42 p.
9. Vollarreal, Raul, 1969. Consideraciones sobre un programa de Mejoramiento de
especies Forestales en Mexico. Degree thesis. Escuela nacional de Agricultura,
Chapingo, Mexico, 128 p.
10. Wright, J.W., 1976. Introduction to Forest Genetics. Academic Press, New York 463 pp.
Annex 1
EXAMPLES OF SELECTION CRITERIA;
SELECTION CRITERIA FOR SEED STANDS
I- Straightness of trunk
1. Straight
2. Slightly crooked
3. Very crooked
II. Forking
1. No forking
2. Forked in upper third
3. Forked in middle or lower third
III. Defects (particularly for Pinus)
1. No defects
2. Recuperated visible defect
3. Average to well-developed defect
* v ' Flowering and fruiting
1. Fruit
2. Flowers only
3. Neither flowers nor fruit
The first three characteristics are determinant: the order of preference is:
111, 211, 112, 121, 122, 221, 222.
Class 3 will only be included in extreme situations, and would in any case
indicate a fairly bad stand.
Flowering and fruiting will serve as a basis for selection between trees in the same
category, when the trees to be left or eliminated are marked.
Data on d.b.h., height and any other obvious features will also be taken.
- 129 -
Annex 2
EXAMPLES OF SELECTION CRITERIA;
SELECTION CRITERIA FOR SEED ORCHARDS
A candidate tree and 4 neighbouring trees to be used for comparison are located.
1. Vigour (basal area or volume)
3 points will be awarded for each 10 percent of superiority of the candidate over
the mean of the neighbouring trees, up to a maximum of 30 points (100% superiority).
The same amount will be deducted if the candidate is inferior to the mean.
2. Trunk
Categories: 1. Straight
2. Slightly crooked
3. Very crooked (only for neighbours)
1 point will be awarded if the candidate is in the best category and 1 additional
point for each 0.25 of superiority over the mean of the neighbours.
The same amount will be deducted for inferiority to the mean.
Maximum number of points: 9.
3. Quantity of verticils
Points will be awarded as follows:
1 point if the candidate is better than the worst of the neighbours but is below the
mean.
2 points if it is equal to the mean.
3 points if it is better than the mean but is inferior to the best neighbour.
4 points if it is equal to the best neighbour.
6 points if it is better than the best neighbour.
Maximum number of points: 6.
4. Average number of branches in the first 3 verticils
The same as for total number of verticils.
Maximum number of points: 6.
5. Crown characteristics
Categories: 1. Live branches only in upper third.
2. Live branches as far as middle third.
3. Live branches as far as lower third.
Scoring as for trunk
Maximum number of points: 9.
- 130 -
6. Diameter of branches
The thickest branch (live or dead) up to the middle of the trunk will be taken.
Categories: 1. No branches
2. Thin branches (up to 1/10 of d.b.h.)
3. Medium branches (up to 1/4 of d.b.h.)
4. Thick branches (more than 1/4 of d.b.h.)
Scoring as for trunk
Maximum numbers of points: 13.
7. Angle of branches
Categories: 1. 90 or more with respect to the trunk above the branch
2. Between 90 and 45
3. Less than 45
Scoring as for trunk
Maximum number of points: 9.
8. Attacks
- 5 points:: healthy tree in a stand that shows slight evidence of attacks
- 10 points: healthy tree in a stand considerably infested (more than 30%)
Basic conditions
1. A candidate tree must not be worse than the worst of its neighbours in any
characteristic.
2. A candidate tree must not be below the mean of its neighbours in more than 2
characteristics (negative values).
3. No tree showing signs of attack by insects or diseases will be considered for
selection.
Contribution of characteristics
Maximum value of plus tree - 92 POINTS
Vigour 32,67. Crown 9.8%
Trunk 9.8% Diameter of branches 14.1%
Verticils 6.5% Angle of branches 9.8%
Branches 6.5% Attacks 10.9%
- 131 -
QUANTITATIVE GENETICS :
GENERAL PRINCIPLES AND PRACTICAL APPLICATION
IN FOREST TREE IMPROVEMENT
BJerne Ditlevsen
National Forestry Service, Denmark
CONTENTS Page
Introduction 131
Quantitative genetics 131
Genotypic and phenotypic values 132
Mean effect of a gene and reproductive value 133
Deviations due to dominance and epistasis 133
Variation of quantitative characters ... ...
Concordance "between related individuals
Heritability 136
General and specific combining ability 137
Selection and genetic gains 138
Bibliography 139
INTRODUCTION
Quantitative genetics deals with the hereditary transmission of those differences
among individuals that may be called quantitative, or differences of degree, unlike
Mendelian genetics, which deals with differences of a qualitative nature (Strickberger,
1968).
The basic genetic functions are the same in quantitative and Mendelian genetics,
but in quairtitative genetics the differences are the result of differences of genes at
many loci, while differences in Mendelian genetics are due to differences in genes at
one or more loci, the different types appearing- in specific proportion (Falconer, 196*0.
QUANTITATIVE GENETICS
Most of the characters which are of interest in connection with forest tree impro-
vement are controlled by a number of additive effects, genes, i.e. such typical quantitative
characters as height, diameter, shape of the trunk and volumetric weight. Studies of
the heredity of these characters must be made on populations, not individuals (Wright,
19T6).
- 132 -
Analyses and studies of quantitative characters can employ a number of different
statistical methods.
Analyses of quantitative characters study the concordances between related indivi-
duals, and one purpose of the analysis is to forecast the effect of a particular selection
and to consider how it should be done.
As mentioned above, statistical methods are essential, in studies of quantitative
characters as are such magnitudes as average and variances.
One important task will be to break down the phenotypic variation ipto its causal
components suchi as external effects and genetic effects. It is also important to break
down the genetic component into additive, dominant and epistatic components (Falconer,
196U ; Shepherd, 1977).
GENOTYPIC AND PHEEOTYPIC VALUES OF THE POPULATION
The genetic composition of a population is often expressed in the form of gene
and genotype frequencies. To find a relation between gene frequencies and the quantitative
differences evident in measurable characters, we may introduce a new term, the term "value",
expressed in the units in which the character is measured (Falconer, 196U).
The following terms are used:
P * The phenotypic value of a given character.
G = The genotypic value. "Genotype" means the complete genetic make up of an
individual.
E * The deviation due to environment, or to all the nongenetic circumstances
influencing the phenotypic value.
The relation among the values may be expressed as: P g G + E.
For the population as a whole, we have:
The mean value m (E) or
m(E)
m (P)
m (G)
The above may be illustrated by a simple example in which the characters are deter-
mined by a single locus with two alleles Al and A2. The individual genotype is given
a value of -a, d and a respectively, as shown in the following equation:
Va
Va
Vi
Genotype
The genotypic value
d * degree of dominance
In the above example, the heterozygote A1A2 does not appear as the mean value of
A1A1 and A2A2, but as the value d, which is called degree of dominance.
If we imagine that the genes Al and A2 appear in a balanced population with the
frequencies of p and q respectively, the genotypic value of the population may be calcu-
lated as in the following table:
T
f Genotype
Frequency Value
Mean value
Vi
p -fa
P 2 .a
Va
2.p.q d
2pq . d
Va
q -a
q . (-a)
TOTAL
1
a(p-q) + 2pqd
- 133 -
The value a(p-q) + 2dpq is the genotypic value of the population insofar as it
refers to this locus only, and at the same time the phenotypic value, since the sum of
population deviations due to environment is zero.
The mean of the population insofar as it refers to all those loci influencing the
character in question (a quantitative polygenic character) is:
M = I a(p-q) + 2 pqd
loci loci
Mean effect of a gene and reproductive value
In the study of the transfer of a "value" of the parents to their descendants,
we cannot use the genotypic value alone, since the parents transmit not their genotype
but their genes to the next generation; instead, the term "mean effect" is used. It is
defined as follows: the mean deviation from the average of the population obtained by
individuals receiving the gene in question (assuming that the allele gene is taken at
random from the population).
The mean effect will therefore depend both on the gene in question and on the
population in question.
Based on the term "mean effect", the reproductive value (A) of an individual may
be defined as follows:
1. The sum of the mean effects of the genes (the genes controlling the character
in question).
The reproductive value is, by virtue of the definition quoted, a theoretical
magnitude. Another, more practical, definition is the following:
2. Twice the mean value of the progeny (measured in deviations from the average of
the population), when the individual has paired at random with a series of
members of the population.
Reproductive value is also called additive genotype.
In a balanced population, we find that the mean of the reproductive values of the
individuals is zero.
The reproductive value of an individual in a population may be calculated on the
basis of the mean value of its offspring after free pollination. According to defini-
tion number 2, the reproductive value may be calculated as follows:
A = 2 x (average of population - average of progeny).
The term "reproductive value", as an expression of that part of the hereditary
mass that can be transmitted in an additive way to future generations, is very important
in forest tree improvement.
Deviations due to dominance and epistasis
In the preceding paragraph on the individual, we mentioned the portion of geno-
typic value that can be transmitted in an additive way to progeny. The remaining portion
of the value of the individual consists of a non-additive gene effect between .the two alleles
of a locus, and is called dominance (D). We therefore have:
- 131* -
For a balanced population we find that the mean of deviation due to dominance is
zero, or
m (D)
If we study a polygenic character, there may also be effects of non-additive
genes among loci. This phenomenon is called epistasis (I) and means that the genotypic
value of polygenic characters may be broken down as follows:
A + D + I.
VARIATION OF QUANTITATIVE CHARACTERS
In the previous paragraph, we discussed characters and values in individuals;
we now discuss the appearance of characters and values in the population. To this end,
let us study first of all the variation of the characters and values and the relation
between them*
The values of the individuals and the variances of the population (in the form
of variance components) may be set out as follows:
P * G + E V p = V r + V_
i P iG E
E 'V* V D* V I* V E
VP Phenotypic variance
VG = Genotypic variance
VA = Additive variance
VD = Dominant variance
VI = Epistatic variance
VE - Variance due to the environment
The simple breakdown of phenotypic variance into the above-mentioned components
of variance is based on the assumption that G and E axe independent of each other and
that A,D and I are independent of one another,
Variances can be used to express the relative importance of a cause of variance
such as the relation between the variance components in question and the phenotypic
variance. The importance of the genotype can, for example, be expressed as the ratio
VG.
VP
The components of genotypic variance and variance due to the environment cannot
be determined directly from observation of a natural population. On the other hand,
the above-mentioned components can be determined by trials or in experimental popula-
tions when the parents are known.
Concordance between related individuals
The components of variance in the previous paragraph can be traced to variations
and therefore we call these components causal variance components.
On the contrary, if we wish to measure the degree of relationship between rela-
ted individuals, another distinct division must be made in the phenotypic variance, the
grouping of individuals into families. These components can be directly estimated on
the basis of phenotypic values, and therefore we call them observed variance components.
To avoid confusion, we will use V for the causal components, <r for the observed
components.
We discuss below the four most usual kinds of relationship, which are:
1. The progeny and one of the parents (the other taken at random from the popula-
tion);
2. The progeny and both known parents;
- 135 -
3- Half-sibs
U.-Full sibs.
The concordance between parents and their offspring can be indicated by the incli-
nation (b) of the regression of parent-children:
Offspring (0)
coefficient of inclination
Parents (P)
b (coefficient of inclination) is estimated as:
P = parents SP sum of pruducts
= offspring SS sum of squares
Concordance among siblings can be indicated by size:
2 2
or or _, Variance among groups of siblings
_ & B
o* B + a, cr s Variance among individuals within groups of
siblings
Now we want to link the observed components with the causal components, i.e. to
find the ratio between the observed values of b and t and the causal components VA and VP.
It is possible to deduce these ratios for the above mentioned relationships. We
give one example below; the relationship of progeny and one of the known parents.
The concordance is indicated by the coefficient b OV PO
According to definition 2 of the reproductive value of an individual the mean (m)
progeny is equal to 1/2A,
or the co-variance must be calculated between one of the parents (with the value - G)
and the mean of the progeny (with the. value 1/2 A).
Cov po = Cov(G, 1/2 A) Cov(A +D t 1/2 A)
SP po ((A + D) (1/2 A) = 1/2$AD + 1/2(A*
Assuming that A and D are indi pendent, the sum of their product is A . D 0.
Cov
C V PO
136 -
The genetic interpretation of the other relationship may also be deduced. The
results are summarized in the following table:
Relationships Concordance observed Genetic interpretation
Offspring and one b 1/2
parent
VA
Offspring and the b ~j
two known parents v ^
t W
PMll sibs t 1/2VA^1AVD
HERITABILITY
Hhe concept of heritability of a character means the proportion of total variance
which is due to the action of genes.
Two concepts of heritability are used, which are defined as follows:
2 VG
1. Heritability in its widest sense h b g * ^
2 VA
2. Heritability in its narrower sense h * =:
n. s v*
It is important to note that heritability is not any fixed magnitude, but depends
both on the population and on the environment. Indication of heritability should always,
therefore, be accompanied by an indication of the environment of the components.
As a general rule, we may say that the herit abilities of current stands are relati-
vely low, due to the fact that the environment has a relatively great impact on the pheno-
typic value of the individual. However, heritability can be fairly high in well-planned
trials.
As mentioned above, a genetic interpretation may be made of the coefficients of
regression observed (parent -off spring) or coefficients between classes (between siblings).
Likewise, the coefficients observed can be used as estimates of heritability. We give
below a summary of the relations between the coefficients and h2 n.s. in the four fol-
lowing relationships:
2
1. Offspring and one parent b * 1/2 h
2
2. Offspring and both parents b = h
3. Half-sibs t * lA h 2
k. Pull sibs t 1/2 h 2
normally, heritability is linked to individuals, but both family heritability and
heritability within the family can be estimated from the guidelines described at the
beginning of this section. Family herit abilities are higher than the corresponding
individual herit abilities since, using average values (average of family), an important
reduction may be achieved in the environmental effect component.
We indicate below how heritability may be calculated on the basis of the results
of a progeny test. The test includes groups of half-sibs and is arranged as a random
block design. The components of variance described in the following diagram of variance
analysis are used to make the calculation:
- 137 -
Source of variation
df
variance components
Progeny (d)
n-1
^ ,4
Blocks (b)
r-1
o* *t" n. or
b
Residual
(n-D(r-l)
0?
Total
rn-1
With groups of half-sibs, the variance between the groups can be interpreted as
lA VA, or:
lA VA; VA
VE =6 /r
If ve assume that there is no dominance, the phenotypic variance could be written
as follows : VF * VA + VE.
2
n.s.
VA
VP
VA
VA+VE
GENERAL AND SPECIFIC. COMBINING ABILITY
The variance between families was formerly divided into genetic components called
causal components. However, the variance between families can also be divided into the
two following components, which can be observed:
1. General combining ability g. c. a.
2. Specific combining ability s. c. a.
The g.c.a. of an individual is defined as the average offspring resulting from
the crossing with various other individuals. The s.c.a. of two specific individuals is
defined as the deviation of their offspring from the average g.c.a. of the two indivi-
duals.
The average offspring between two individuals x and y can therefore be described
as follows:
GCA,
Consequently, the variance between crossings may be analysed in two components:
variance in the general combination capacity and variance in the specific combination
capacity. The latter appears in the statistical analysis as the interaction component.
From definition 2 of reproductive value it may be seen that the g.c.a. is equal
to half the reproductive value of the individual.
The specific combination capacity depends exclusively on non-additive genetic
variance.
A knowledge of g.c.a. and s.c.a. is important for the selection and combination of
material for future seed production.
G.c.a. and s.c.a. can be calculated on the basis of crossing designs that will be
described in more detail in a later article, and we give below only one example of the
calculation of g.c.a. and s.c.a.
-138-
1. Mean values of four progenies:
Parents
1
2
Average
3
20
60
1*0
J
30
90
60
Average
i
25
75
L
50
2. Deviation from the average total of the progeny and calculation of the g.c.a.
i
Parents
i
2
g.c.a.
3
-30
10
-10
1*
-20
1*0
10
g.c.a.
-25
25
3. Calculation of the s.c.a. through the equation
B.c.a. 13 * -30 -(-25) - (-10) * 5
s.c.a.
23
10 - 25 - (-10) * -5
-20 -(-25) - 10 = -5
1*0 - 25 - 10 * 5
SELECTION AND GENETIC GAINS
Some of the above mentioned genetic parameters, particularly heritability, are
very important for selection and the estimation of genetic gains through selection.
The strength of selection in a population may be expressed as the selection diffe-
rential S which indicates the difference between the mean of the population and the mean
of the selectedgfraction, or as intensity of selection i ^hich is expressed through the
* *- - being equal to the phenotypic dispersion of the population.
formula i
If we select the best individuals in a population, we can express the genetic
gains (R) as the displacement of the mean among the progeny of the selected individuals
and the entire population.
The relation is illustrated in the following figure:
- 139 -
Offspring (0)
s b = coefficient of slope
Average of the tvo
parents (P)
The slope of regression which expresses the relation between S and R is equal to
heritability.
P g p
The coefficient of regression b = h -r-, hence R = h S.
n. s R n.s.
The selection differential S. depends on the dispersion of the population, and it is
therefore a good idea to standardize it by dividing it by phenotypic dispersion, as in
the case of i.
Using i instead of S the following function is obtained:
R = i.hj ap
n.s*
BIBLIOGRAPHY
Falconer, D.S.
Shepherd, K.R.
1977
Strickberger, M.W.
1968
Wright, J.W.
1976
Introduction to Quantitative Genetics, Oliver and Boyd, Edinburgh
General Genetics and Inheritance, International Training Course in
Forest Tree Breeding, Canberra, Australia.
Genetics, The Macmillan Company, New York, U.S.A.
Introduction to Forest Genetics, Academic Press, New York, U.S.A.
- 140 -
VEGETATIVE PROPAGATION METHODS
Marcelino Quijada R.
Institute de Silvicultura
Unlversidad de Los Andes
Her Ida, Venezuela
CONTENTS
Page
Definition 140
Practical use 140
General terminology 141
Methods 141
Cuttings 141
Grafting 142
Layering 143
Tissue cultures 143
Incompatibility 143
Bibliography 145
DEFINITION
Plants can be propagated sexually or asexually. The basic difference is the pro- *
cess of fertilization that occurs in the first but not in the second method.
Asexual propagation, in turn, takes two forms: apomixis and vegetative propagation.
Apomixis is propagation through the development of one of the gametes (the ovule) without
fertilization, or of a cell without reduction, to form seeds or seed-like organs.
Vegetative propagation is propagation from well-differentiated vegetative parts.
PRACTICAL USE
Practical use of vegetative propagation methods is based on two biological consi-
derations:
a) Maintenance of the physiological condition of the parent tree in the propagated
part.
b) Maintenance of genetic constancy. That is to say, the part propagated is gene-
tically identical to the original individual.
Vegetative propagation has been widely used in breeding for, among other things:
a) the establishment of clonal seed orchards;
- 141 -
b) the establishment of clone banks, in which controlled pollination is effected,
owing to the possibility it affords for obtaining flowers at low height;
c) the propagation of special breeding material: exceptional hybrids (e.g. hetero-
tics) that are lost through sexual reproduction, sterile hybrids, etc.;
d) the propagation of selected plants on a large scale.
Its usefulness depends on, among other factors:
a) The ease with which the species can be manipulated. Many species are difficult
to propagate vegetatively, others are extremely easy. This often affects
production costs, both for the establishment of orchards and for large-scale
production of plantation material.
b) The extent to which development of the parts propagated can be controlled.
In some cases the phenomenon of topophysis occurs: the development of the
propagated part is influenced by the part of the tree from which it comes,
e.g. lateral branches sometimes tend to grow in a horizontal direction.
Another phenomenon which affects development is graft incompatibility: in grafting
the scion may be rejected, sometimes after a year or more.
GENERAL TERMINOLOGY
The original tree from which the parts to be propagated are taken is known as the
ortet. Each part already propagated is a ramet . A set of ramets from the same ortet is a
clone. Differences observed between trees propagated by the same method is called clonal
variation.
METHODS
There are three main methods of vegetative propagation: cuttings, layering and
grafting.
Cuttings are sections taken from the tree and put to root in an appropriate
medium.
Grafts are plants obtained by fusing a part from the tree to be propagated with
another part which has its own root.
Layers are sections of the tree which are induced to root and then separated
from the tree.
For breeding purposes the methods most used are cuttings and grafts. Layering
can be of help in special cases as a transitory measure.
Cuttings
Cuttings may be of wood or leaves. The first come from branches or the trunk and
are more important in forestry. Leaf cuttings are more common for ornamental plants
particularly those with fleshy leaves.
It is the most economical method of vegetative propagation in species with good
reserves of aqueous tissue, such as Bombacaceae.
Rooting depends on, among other things, the age of the plant, the condition of
the cutting, the time of collection, the rooting medium and special treatments.
- 142 -
Tree age is difficult to determine in natural tropical forests, buy the dimensions
of the tree, in particular the diameter, can be used as a guide. Very stout trees, more
than 50 cm. in diameter, are usually old and therefore do not root easily; when rooting
is achieved, the root system often develops poorly. The best propagation is achieved
with young trees, but this often runs counter to specific purposes, such as the production
of flowers, fruit and seeds.
As regards the condition of the cutting, very woody cuttings do not sprout or
root easily. Very herbaceous cuttings tend to be very susceptible to drying out. An
intermediate condition is therefore best, particularly when the cutting is planted out
in the field immediately. A guide for the collection period is the stage of activity
of the plant; the most favourable period is when the buds on the tree are starting their
most active growth. In the low-lying tropical zones of Venezuela this coincides with the
onset of the rainy season (March/April).
The rooting medium must guarantee sufficient but not excessive moisture; this is
normally achieved with a medium, semi-sandy textured soil and adequate atmospheric humi-
dity. This is easy to control under artificial conditions, but in the field it can only
be guaranteed during the rainy season.
As regards treatments, rooting hormones such as indoleacetic, indolebutyric and
naphtalenacetic acid are useful for small-scale work (and even for orchards). For com-
mercial purposes, powdered dusts such as Root one, Hormodin, Hormonagro, etc. or pure
solutions of 100 to 1 000 p. p.m. are recommended.
After the cutting has been planted, the upper surface, if it is the site of the
cut, should be protected with a substance to reduce evapotranspiration.
Grafting
This involves the removal of a vegetative part from the parent tree (the scion)
and its attachment to a part with its own root (the stock) so that the tissues fuse.
Depending on the taxonomic relations between stock and scion, we have hetero-
plastic grafts if they belong to different species (e.g. Cedrela odorata on . angusti-
folia) and homoplastic grafts if they belong to the same species. Homoplastic grafts
are called autoplastic if stock and scion have the same genotype.
According to the position of the scion, we have top and side grafts, depending
on whether the scion is inserted into the top or the side of the stock.
Top-grafting entails beheading the stock, which in many cases results in a dras-
tic reduction in the plant's foliage. This technique is used for many hardwood species.
Side grafting makes it possible to maintain adequate foliage on the stock, and
has been much used for conifers, where the moisture factor is very important.
Diagrams of some grafts commonly used for forest trees are given in Annex 1. A
good, step-by-step description of grafting and layering with special reference to pines,
is given in Dorman (1976).
The cuts should be made with grafting knives, which should be kept clean and
sharp, in order to ensure a clean cut which will not be a focus for infection. These
knives also have a metal or hard rubber spur to separate the bark in budding.
Cutting must be rapid and uniform for a clean, even cut.
The scion should be bound to the stock with a strong but slightly flexible mate-
rial. A semi-elastic plastic strip called grafting tape is usually used. It may also
contain a fungicidal compound to prevent attacks by fungi.
- 143 -
Outdoor grafts must be effectively protected to prevent the scion drying out. A
double covering is often used, the inside layer consists of a plastic bag which helps
to retain moisture, and the outer of cloth or non-transparent paper which protects the
graft from the direct rays of the sun. As a precaution, it is advisable to effect the
graft on a cloudy, slightly humid day.
In selecting the scion the main consideration is its stage of growth. Buds that
are in full activity should be sought; this can be checked by their size and by the time
of collection (little differentiation and collection at the onset of vegetative growth).
The scion can be the bud itself with part of the bark to attach it to the stock,
or it can be a piece of the branch with one or more buds. In the latter case the piece
of branch must be neither too soft nor too hard, a second-year growth is ideal.
Care must also be taken to see that the scion is kept in a cool, damp place for
no more than two days, if not used immediately.
The stock must be young and healthy. The main guide is compatible thickness of
scion and stock. Diameters of between 0.5 and 2 cm. are preferable to ensure good adhesion.
Layering
It may often be difficult to propagate a species by cuttings, and grafting may
present problems. One possible solution is layering. The procedure consists of making
a wound in a section of the tree and covering this with a medium which retains moisture
(moss, earth, etc.). Healing causes the formation of a callus from which adventitious
roots may grow.
This process is helped along by hormone compounds similar to those used for cut-
tings.
The most common type of layering is aerial layering in which the covering is a
medium other than soil and the operation takes place rather high off the ground. When
long, low, flexible branches are available, these can be inserted into the soil; we then
have ground layering. Usually the part propagated is a branch, but in some species (e.g.
Platymscium sp.) interference with the roots can cause them to sprout, constituting a
form of layering in that they do not separate from the stem.
The recommended size is as for cuttings; branches with a diameter of 1 to 2 cm.
One of the biggest problems of this method is transplanting, which often results
in considerable losses. It is recommended that the first transplant be made in a fairly
loose (sandy) medium to encourage root adaptation. The plant can then be transplanted
into a more compact medium.
Tissue culture
Another form of vegetative propagation consists of tissue cultures, (cultures of
cells with the potential mitotic activity, in an appropriate medium under aseptic condi-
tions). The method has been tried with varying degrees of success in gymnosperms and
angiosperms, and everything seems to indicate it is feasible for many forest species.
The great advantage is that only a very small portion of material is needed. The main
disadvantage is the very controlled work conditions required and the same limitations as
for cuttings, grafts, etc. (e.g. genotyplc variation in ease of propagation).
VEGETATIVE INCOMPATIBILITY
One of the main problems in grafting is incompatibility which produces rejection of
the scion by the stock.
Rejection simply indicates (of ten) that the cut areas have not been fitted together
well (e.g. uneven cuts minimize points of contact). Sometimes, however, rejection is
- 144 -
an indication of conditions inherent in genetic somatic differences between the tissues.
This phenomenum may occur at an early stage, making it necessary to effect more grafts
to fulfill a given quota (e.g. clone(s), problems with clone(s)). A greater problem is
that of delayed incompatibility which may emerge one or more years after the grafts have
been established in the field, thus leading to greater losses, e.g. losses of producti-
vity in the case of seed orchards.
In many species external signs have been detected which indicate incompatibility.
These indlude: a) consistent early defects in specific trees, b) unequal rate of growth
of scion and stock, c) excessive growth in, above or below the juncture zone, and d)
abnormalities in colouring and leaf development of the scion.
The last three are obvious signs of delayed incompatibility which appear between
six months to ten years after grafting.
This phenomenon may occur not only in heteroplastic grafts, where it might be
expected because of taxonomic differences, but also in homoplastic grafts, i.e. within
the same species. In the latter instance, it has more to do with differences in prove-
nance or geographic source. One way of controlling this phenomenon is to seek the great-
est possible affinity between the plant serving as stock and the tree from which the
scione are taken, to ensure optimum histological affinity.
It is to be noted that there is no best method of grafting: success often depends
on the skill of the grafter using a specific method. Various tests have shown different
degrees of difficulty in manipulating scion and stock, but the final results have been
practically the same as regards percentage of take, as shown below.
Vegetative propagation characteristics of some forest species in Venezuela^
Species Ease of vegetative propagation
Bombacopsis qulnata VE
Cedrela odorata M - E
Tabebuia rosea M - E
Swietenia macrophylla D
Anacardium excels urn M - D
Pithecelobium saman yp
Podocarpus rospigliosil D - VD
Cordla alliodora M - D
Cordia apurensis M - D
Podocarpus oleifolius M
Hura crepltans M - E
Tectona grandis
Gmelina arborea
Plnus caribaea v. hondureneis M - E
Plnus oocarpa M - E
Pinus radiata M - E
Plnus patula M - E
- 145 -
N.B. VE Very easy, percentage above 80%
E - Easy, average percentages 60%
M Moderately difficult, percentages around 40%
D - Difficult, percentages around 20%
VD - Very difficult, percentages usually 0%
Species between 2 classes show great clonal variability.
BIBLIOGRAPHY
Dorman, K.W. The Genetics and breeding of Southern Pines. U.S. Dept. Agriculture,
1976 Agriculture Handbook no. 471. Washington D.C.
Hartman, H. and D. Kester. Propagacion de Plantas. Companla Editorial Continental,
1967 S.A. , Mexico. 693 pp.
Jett, J.b. Vegetative Propagation.
1969 In Forest Tree Improvement Training Course.
N.C. State University, Raleigh N.C. pp. 217-225
Jett, J.B. Incompatibilty. In Forest Tree Improvement Training Course,
1969 N.C. State University, Raleigh, N.C. pp. 226-230
Koenig, A. and Melchior, G.H. Propagacion Vegetativa en Arboles Fores tales. Proyecto
1978 Investigaciones y Desarrollo Industrial Forestales - COL/74/005. PIF
no. 9. Bogota, Colombia.
New Zealand Forest Service. Special Issue on Vegetative Propagation.
1974 N.Z. J. For. Sc. .4(2): 119 - 458.
Quijada, M. and V. Gutierrez. Estudio sobre la propagacion vegetativa de Especies
1972 Forestales Venezolanas. Rev. For. Ven. 21: 43-56.
The Swedish University of Agricultural Science. Vegetative Propagation of Forest
1977 Trees - Physiology and Practice. Uppsala, Sweden, 159 pp.
Wright, J.W. Introduction to Forest Genetics. Academic Press, New York,
1976 463 pp.
- 146 -
ANNEX 1
FOREST TREE GRAFTING METHODS -
II
In a graft the cambium of the stock and the scion should coincide at least on one
side in order to form the new tissues uniting the two parts rapidly. The cuts must he
flat and smooth so that as few cells as possible are damaged. To obtain these conditions
various grafting methods have been developed, depending on the diameter of scion and
stock. The main grafting methods used for forest trees are the following:
a) Cleft graft
Scion
Bark of scion
Stock |.fifc'f . Bark of stock
Scion
In cleft-grafting the scion cut in the form of a wedge is inserted into a cut made
in the top of the stump. Scion and stock must be of approximately the same diameter. In
simple or radial cleft-grafting, the scion occupies only one side and may have a diameter
inferior to that of the stock.
b) Splice or whip and tongue graft
Scion
Stock
jBjScio;
Stock
Scion and Stock (similar diameters) are joined by means of a long diagonal cut so
that the cut on the stock coincides with the cut on the scion.
J7 From Koenig and Melchior (1978).
c) Veneer-graft
- 147 -
Stock-
Stock Scion
This method is used when the diameters of stock and scion differ considerably. The
stock is prepared by a slanting cut which penetrates as far as part of the xylem, and the
cut piece removed. The scion is prepared so that the cut coincides with that in the stock.
d) Underbark graft
Stock
Scion
This method Is also used for different diameters, only the bark of the stock being
cut. The scion, cut in a triangular shape, is placed in the opening.
b) Budding
Bud shield
Bud
The insertion of buds used in the propagation of fruit trees and roses, called
budding, can also be used for tropical forest trees, if only a small number of scions are
available, or if the methods described above do not prove successful enough.
- 148 -
CONTROLLED CROSSING SYSTEMS AND DESIGNS
B. Ditlevsen
National Forestry Service, Denmark
CONTENTS
Page
Introduction 148
Purpose of controlled crossings 148
Selfing 149
Crossing systems with unknown father 1 50
Free loss of flowers I 50
Poly-crossing 150
Crossing systems with known father 151
Complete diallel crossing plan 152
Modified diallel crossing plan 152
Partial diallel crossing plan 154
Factorial crossing plan 1 55
Single pair pairings 157
Choice of crossing plan 157
Bibliography / 159
INTRODUCTION
One of the most important decisions to be made in those genetic improvement programmes
that include controlled crossings, is the choice of a crossing system. Apart from tree
selection, the choice of a crossing system is the only way in which the tree breeder can
improve a genetically variable population.
We describe below a series of important aspects to consider when choosing a crossing
system. In practice, however, it is often difficult to choose the best system, since there
are usually different competing requirements that cannot always be satisfactorily met by
one 'system alone. The final choice, therefore, will probably be a compromise solution.
Lastly, there may be some practical and economic constraints.
PURPOSE OF CONTROLLED CROSSINGS
As indicated above, the purposes of controlled crossings are many and various, and
we set out below the most important ones for forest tree improvement (Brown, 1977, Roberts,
1969).
1. Determination of general combining ability (g.c.a.). It is usually difficult to
estimate the reproductive value, or general combining ability of an individual
tree, from the phenotypic value of the individual; only studies of its progeny
will provide us with the necessary information on the capacity of the individual
to transmit its good characters to its offspring.
- 149 -
Information on the g.c.a. of individuals can in practice be used for:
a. Choosing the best individuals for seed orchards;
b. Genetic thinning in existing seed orchards;
c. Choosing the parents of the progeny which it is desired to use in future
genetic improvement.
2. Determination of specific combining ability (s.c.a.). Data on specific
combining ability can be used in establishing seed orchards of two
clones; this makes it possible to select for special effect in the
offspring of two individuals.
3. Determination of variance in general and specific combining ability*
4. Estimation of heritability. Knowledge of both the variance in
g.c.a. and s.c.a. and heritability is very important in deciding
for the establishment of the best procedure for genetic improvement
and for optimization of, for example, the number of individuals for
progeny and the number of progeny per clone.
5. Production of material for selection of individuals for the next
feneration of seed orchards. We may assume that the best individual s
n a group of siblings are better than the average of the parents and,
consequently, it is better to select the best individuals within the
best groups of siblings for use in the future seed orchard.
6. Production of material for continued improvement. To raise the
genetic quality still further from the level achieved by a seed
orchard of clones whose progeny has been evaluated, it is necessary
to recombine the hereditary characters through crossings and new
selections. This procedure can be repeated several times before selecting
the individuals that will form part of the future generations of seed
orchards. In this way it is possible to eventually achieve several
combinations of genes that are very seldom found growing wild but that offer
great advantages from the breeding standpoint.
7. Estimation of genetic gains in the first generation of seed orchards and
in successive generations. Decision-makers should be able to get
information on the degree of improvement (genetic gains) achieved in seed
orchards as compared with seeds from stands, and on the genetic gains that
can be expected from the establishment of seed orchards of the next generation.
Lastly controlled crossings could be useful in practical forestry work to obtain
plus material. For example, two clones might have a very high capacity for specific
combination, but without simultaneous flowering periods; they could not therefore,
be used in a seed orchard of two naturally pollinated clones.
The establishment of controlled crossings of forest trees is an expensive
time-consuming operation. In consequence, it is often best to use crossing designs
that meet several of the desired objectives at the same time.
SELF-POLLINATION
Controlled self-pollination is studied mainly in research on selfing. In
seed orchards, selfing clones can produce much selfed seed, which in turn produces
minus plants*
Controlled self-pollination can also form part of an intra-crossing/crossingg
progranme in which attempts are made to produce strongly intra-crossed individuals
which are later crossed with other, also intra-crossed, individuals to achieve a
heterosis effect. The method is better known in agriculture, but it has also been
tested in forest tree improvement (for example, with Larix, see Keiding (1968). The
possibilities of using intra-crossings and crossing* in forest tree breeding are
discussed by Lindgren (1975).
- 150 -
CROSSING SYSTEMS WITH UNKNOWN FATHER
In this paragraph we shall discuss two designs: free loss of flowers and
poly-crossing,
Free loss of flowers
After free loss of flowers the seed is usually collected by one of the two
following methods.
1. Collection of seed from plus trees selected from a stand. The plus tree
has been fertilized by the other trees in the stand, and it may also have
many selfed seeds.
2. Collection of seed from established seed orchards. If the seed orchard or
the plantation is producing large quantities of pollen, it may be assumed that
individual clones in the plantation have been fertilized by the other clones
in the plantation or, in some cases, by self-pollination.
In both cases, the uncertainty of the relationship between the individual seeds
is considerable. In analyses of results of tests and in later interpretation it is
normal to assume that individuals of the progeny after free loss of flowers are half-
sibs t or that they have the same mother but different fathers.
We have already mentioned that in certain cases certain quantities of selfed
seed may be expected, and also, it may be assumed that some of the progeny are full
sibs (that they have a common mother and father).
The uncertainty as regards relationship may lead to erroneous interpretations,
and the most important disadvantage of free loss of flowers is the difficulty in
determining the general combining ability due to the fact that the fathers are unknown
and that they vary.
Poly-crossing
The disadvantages of free loss of flowers, namely the risk of self-pollination
and fertilization by different pollen for different mothers, can be avoided through the
so-called poly-crossing system. In poly-crossing, artificial pollination is effected,
using a mixture of pollen from a large number of fathers (not, however, from the mother's
clone). In this way self-pollination may be avoided, and all the mothers are pollinated
with a constant mixture of pollen.
So many different fathers are involved that specific effects of combination, if
any, are neutralized, and poly-crossing is therefore a good, cheap system for determining
g.c.a. However, it should be mentioned that non-randomized pollination (because the
pollen of some clones is more viable than that of others) can result in erroneous estimates
of g.c.a. (Roberds, 1969).
One characteristic common to free loss of flowers and poly-crossing is that,
since the fathers of the offspring are not known, the possibilities of making a selection
of progeny for use in the next generation of seed seedlings are considerably limited.
On the one hand, it is impossible to select progeny with a good quality father, and
on the other hand, there is the risk of choosing individuals already related to each
other.
Systems in which the fathers are unknown do not allow for evaluation of the possible
specific effects of combination.
Nevertheless, the above-mentioned disadvantages of the unknown fathers system
are offset by the fact that these systems are cheap and simple to implement, and both
the trials and the statistical analysis are easy to carry out.
The following is a diagram of the analysis and interpretation of a poly-crossing
design laid out in a field trial as a random block design. The genetic interpretation
- 151 -
has been based on the assumption that the progeny are half-sibs.
Variance Analysis
Source of variation
df
MS
expected MS
Blocks
b-1
Families
a-1
M l
j*^*;
Error
(a-1) (b-1)
M 2
**+"**
Within plots
a-b-(w-l)
M 3
i
Genetic interpretation
a -
o
"n.s.
1/4
V P -
-- .
v. ;
Variance due to families
Environmental variance among plots
Environmental variance among trees in the plots
Additive variance
Dominant variance
Environmental variance within plots
Phenotypic variance
Genotypic variance
CROSSING SYSTEM WITH KNOWN FATHER
In this group of crossing systems the offspring are full sibs, i.e., the individual
offspring have a common mother and father. Therefore, as a general rule, we may expect
to have better information from progeny trials than from systems where the father is
unknown. However, systems where both the mother and the father are controlled are
more difficult, so various fairly complete designs have been developed; they are described
In detail below.
- 152 -
Complete diallel crossing plan
Figure 1 shows a complete diallel crossing plan.
Fathers 12345678910
1
V N
S?c
X
X
X
X
X
X
X
X
2
x>
s?
X
X
X
X
X
X
X
3
X
x>
V
XX
X
X
X
X
X
X
4
X
X
xS
V s
S, x
X
X
X
X
X
5
X
X
X
xS
V
v*
X
X
X
X
6
X
X
X
X
v x
X
X
X
7
X
X
X
X
X
xS
X.
s?
X
X
8
X
X
X
X
X
X
/x\
"X
X
9
X
X
X
X
X
X
X
xS
^xS
V.
10
X
X
X
X
X
X
X
X
xS
v X
Fig. 1. Complete diallel crossing plan. Self-pollination is indicated in the figure.
This crossing plan is the best one, since it includes all crossing possibilities
and provides the most complete information on the genetic characters of the clones
studied.
This design can provide information on the general and specific effect of
combination and on its variances. The material also constitutes the best starting point
for the selection of superior individuals or pairs of clones suitable for two-clone
plantations.
For practical purposes this design is, unfortunately, very difficult to use.
In the first place, it usually yields clones that produce too small a quantity of male
or female flowers; however the most important disadvantage is an economic one. A complete
diallel crossing plan with, for example, 20 clones, will require 400 controlled crossings,
or 380 if self-pollination is avoided. However, it is not very realistic to spend
so much on each clone, and a cheaper design, such as the one described below, is more
generally used.
Modified diallel crossing plan
One way of shrinking the crossing plan, and thus making it cheaper, is to omit
reciprocal crossings and self-pollinations, as shown in Figure 2.
- 153 -
Parents
1
2
3
4
5
6
7
8
9
10
10
x
X X
XXX
X X X X
X X X X X
X X X X X X
X X X X X X X
xxxxxxxx
xxxxxxxxx
Fig. 2. Modified diallel crossing plan.
This design provides roughly the same information as the full diallel crossing
plan, but the limited material in the plan does not, of course, guarantee the same accuracy
in trials and determination of parameters.
Below is a diagram of the statistical analysis and the genetic interpretation of
a modified half diallel crossing plan in a random block design.
Variance analysis
Source of variation
df
MS
Expected MS
Blocks
b-1
Families
a-1
GCA
n-1
M l
a w * ^e * wb s * wb(n ~ 2)0 2
SCA
n(n-3)/2
M 2
a 2 + wo 2 -i- wba 2
we s
Error
n(n-l)(b-l)/2
M 3
a 2 * wo 2
w e
Within plots
nb(n-l) (w-l)/2
M 4
a 2
W
GCA - General combining ability
SCA - Specific combining ability
Genetic Interpretation
2
1/4 v_
1/4 V D ,
4 . a'
- 154 -
Partial diallel crossing plan
This design may be distinguished from the complete and modified diallel plan in
so far as a clone has not been crossed with all the other clones. The design can have
a series of different conformations (Braaten, 1965) and in Figure 3 we indicate two
different types.
Parents
10
1
X X X X
2
XXX
3
X X
4
X
5
6
X X X X
7
XXX
8
X X
9
X
10
Disconnected partial diallel design.
Parents
1
2
3
4
5
6
7
8
9
10
11
8
10 11
x x
X X
X X
X X
X X
X
X
X X
X
X
X X
Brown's partial diallel design (Kempt borne and Cur now, 1961)
Fig. 3 Partial diallel crossing plans.
As may be expected, these designs are less efficient than the full design and
the modified diallel design. However, to make up for this, a large number of offspring
can be tested at a fairly low cost. Unfortunately, missing values, which cannot be
avoided in large-scale crossing plans, considerably complicate the work of calculation.
As illustrated in Figure 3, there are various different designs in the group of
partial diallel designs, and we give below an analysis of one of the most frequent, the
disconnected partial diallel design.
- 155 -
Variance analysis
Source of variation
df
MS
Expected MS
Blocks
b-1
Diallels
d-1
Block x diallel
(b-1) (d-1)
Families within
diallels
GCA
d.(c-l)
d.(p-l)
M l
222 2
w * WQ e * bwQ s * bw < n ~ 2)a
SCA
dp(p-3)/2
M 2
a 2 + wa 2 + bwa 2
we s
Error
d( b -l,(c-l,
M 3
w e
Within plots
dbc-(w-l)
M 4
a 2
W
Genetic interpretation
M. - M,
9 bw(n-2)
, 2 - M 2" M 3
8 bw
.2 M 3- M 4
1/4
1/4
v^ - a 2 - a
G g s
^
Ew
rr 4- rr
Factorial crossing plan
In this design a number of mother clones are crossed with the same number of father
clones. Often there is a small number of fathers, also called common testers. This
design can also be considered a sort of full diallel design including all the combinations
of one group of mothers and another group of fathers. See Figure 4.
- 156 -
Parents
1
2
3
4
5
X
X
X
X
6
X
X
X
X
7
X
X
X
X
8
X
X
X
X
9
X
X
X
X
10
X
X
X
X
11
X
X
X
X
12
X
X
X
X
Fig. 4 Factorial crossing plan.
The design is widely used in the United States under the name of North Carolina
II. Normally four different fathers are used for the crossing plan, but as a general
rale, the number of fathers in the plan must depend on the magnitude of the specific
combining effects. If major specific combining effects are likely, the estimation of
g.c.a. of the mothers may turn out to be highly erroneous if too few father trees are
used.
Since the design often includes very few fathers and since the same clone does
not appear as mother and father, it is difficult to compare the general g.c.a.'s of each
parent .
It is also difficult to make a progeny selection for the next generation of seed
orchards, since selected individuals, especially when only a few fathers are used, are
very often already related to each other.
One advantage of this design is that its implementation is simple and therefore
relatively cheap; at the same time it makes analysis of the results easier. We give
below a diagrammatic summary of a factoral design.
Variance analysis
Source of variation
df
MS
Expected MS
Blocks
b-1
Fathers
Mothers
m-1
f-1
M l
M 2
a w * ^e * ^^mf * bwf a m
t
Fatherx x mothers
,.-!, ,-U
M 3
i**:**i
Within plots
<*-!) (b-1)
M 4
i*-i
hBf (*-!)
M 5
a 2
W
- 157 -
Genetic interpretation
- 1/4V A ' V A
V A
a 2 - M. - v_ - a 2 - a 2 - a 2 + v_ v ft a 2 + a 2 + a 2 . + a 2 + a 2
w 5 Gm f mEw;P mfmfew
Single pair pairings
In this system each clone is involved only once as mother or father.
The system works particularly well if the objective is to produce a population
for selection of individuals for new seed orchards or for use in continuous breeding
work.
Another advantage of the use of single pair pairing is that a large number of
clones may be tested under the same plan and normally it is very expensive to produce
offspring on the basis of controlled crossing.
On the other hand, the possibilities of estimating the g.c.a. and the variance
of the general and specific combining abilities are normally not very good. If the
specific combining effects are not important, a quick estimate of g.c.a. may, however, be
made in order to set aside the worst of the clones tested. If, on the other hand,
the specific combining effects are very important, this design may be used to select
the best combinations of clones for use in two-clone plantations.
CHOICE OF CROSSING PLAN
The definitive choice of a crossing plan will depend mainly on the purpose of
the controlled crossings, as described in the first paragraph of this article. The
advantages and disadvantages of the different crossing systems and designs have been
briefly discussed in the paragraphs above, and we provide a summary of all the most
important aspects in Table 1. Lindgren (1977) and Buijtenen (1976) have given a more
detailed analysis of the different designs.
Apart from the problems discussed, which are important in the choice of a crossing
plan, a series of problems of a practical nature may arise in the implementation of
large series of crossings; for example the differences between clones insofar as their
time of flowering is concerned.
We may also mention problems related to the transfer of the crossing plan to a
field design. If possible, a simple, sturdy test design, such as the random block,
should be chosen* With many different offspring it might, however, be necessary to
use an incomplete design. See Braaten (1965).
- 158 -
Table 1. Comparison of crossing plans
Crossing
System
Determination
of CCA
Selection of
plus trees
Costs
Determination of
variance of GCA
and SCA
Free loss
of flowers
Fair
Possible, but
inefficient
Low
Difficult
Poly-
crossing
Very good
Very little and
only if the
depression of
Very low
Good determination
of the variance of
GCA
intra-crossing
is reduced
Full
diallel
Excellent
Excellent
Very high
(impossible
where there is a
Excellent
large number of
clones)
Modified
diallel
Excellent
Excellent
Very high
(impossible
where there is a
Very good
large number of
clones)
Partial
diallel
Good
Very good
Fair
Good determination
of the variance of
GCA. The variance of
SCA can be determinei
but it is difficult
from the point of
view of data process*
ing
Factorial
Good
Only in a few
cases and
Fair
Good
where intra-
crossing
depression is
low
Single
Bad
Good
Very low
Bad
pair
pairings
GCA - General combining ability; SCA - Specific combining ability
- 159 -
BIBLIOGRAPHY
Brown, A.G. , 1977. The Strategy of Tree Improvement. International Training Course in
Forest Tree Breeding, Canberra, Australia.
Braaten, M.O., 1965. The Union of Partial Diallel Mating Designs and Incomplete Block
Environmental Designs. Inst. of Statistics. Mimeograph Series No. 432.
Keiding, H. , 1968. Preliminary Investigations of Inbreeding and Outcrossing in Larch.
Silvae Genetica 18(5).
Kempthorne, 0. and Curnow, R.N., 1961. The Partial Diallel Cross, Biometrics 17.
Lindgren, D. , 1975. Use of Selfed Material in Forest Tree Improvement. Dept. of Forest
Genetics, Research Notes 15.
Lundgren, D. , 1977. Genetic Gain by Progeny Testing as a Function of Mating Design and
Cost. Third World Consultation on Forest Tree Breeding, Canberra, Australia.
Ro herds, J.H., 1969. Progeny Testing in Forest Tree Breeding, FAO-North Carolina State
Forest Tree Improvement Training Centre.
Van Buijtenen, J.P., 1976. Mating Designs, IUFRO Joint Meeting on Advanced Generation
Breeding, Bordeaux.
Participants visiting the CVG plantations
of Pinus caribaea in Uverito
- 160 -
SEED ORCHARDS -
W.H. Barrett
Fiplasto S.A., Buenos Aires, ARGENTINA
CONTENTS Page
Introduction 16 1
Concept and design 161
Definitions 161
Historical background 161
Planning and design 162
Clones or seedlings 162
Number of clones or seedlings 162
Initial planting distance 163
Experimental design 163
Vegetative propagation 163
Establishment 1 64
Location 164
Site preparation 1 64
Planting 164
Management and harvesting 165
Cultivation techniques 1 65
Plant cover 165
Fertilization 165
Pruning and thinning 166
Protection . . . 1 66
Flowering, pollination and handling of pollen 166
Harvesting of fruit and seed 1 66
Seed production 167
Seed extraction and utilization ... 167
Keeping records 167
Bibliography 167
\J The text of this article has followed the structure and content of the book, "Seed
Orchards", edited By R. Faulkner (1975) for the IUFRO (International Union of Forestry
Research Organizations) Working Group S 2-03-3, with additional information taken from
lectures given by D.G. Nikles in Kenya in 1974 and A.G. Brown and K.G. Eldridge in
Australia in 1977.
- 161 -
INTRODUCTION
The modern trend of producing wood and wood products from plantation forests implies
the availability of enormous quantities of seed, if extensive areas of forest are to be
planted.
It is sometimes difficult to obtain enough high-quality seed in sufficient quanti-
ties for the increasingly large-scale expansion plans in the different forest regions
of the world.
Moreover, the cost of forestry activities, from the preparation of the soil to the
felling of the trees and extraction of the logs, and also the irrecuperable factor of the
long wait until the trees are ready to harvest means the breeder must know in advance,
that the seed will be good.
Seed orchards ensure a supply which is regulated by the foresters themselves, and
which provides steady, reliable production for their plantation programmes.
The establishment of seed orchards is therefore accepted as a necessary forestry
activity, which should be started as soon as possible in afforestation programmes relying
on seed supplies.
CONCEPT AND DESIGN
Definitions
According to the general definition by Faulkner (1975), seed orchards are the
most important means the tree breeder has for mass-producing seed for large improved
plantations, based on the best selected trees.
The classic definition by Zobel e_t al (1958) states that "A seed orchard is a
plantation of genetically superior trees, isolated to reduce pollination from genetically
inferior outside sources, and intensively managed to produce frequent, abundant, easily
harvested seed crops. It is established by setting out clones (as graftings or cuttings)
or seedling progeny of trees selected for desired characteristics".
In some cases, seed orchards are established to mass-produce seed of some popula-
tion of which it is impossible to obtain seed in adequate quantities, without too much
attention being paid to the genetic superiority of the individuals. For this reason,
the Organization for Economic Cooperation and Development broadens the concept of the
seed orchard for the use of international trade, as follows (Brown and Eldridge, 1977):
"A seed orchard is a plantation of selected clones or progenies which is isolated or
managed to reduce pollination from outside sources, managed to produce frequent,
abundant, easily harvested crops of seed."
HISTORICAL BACKGROUND
Although the concept of a seed orchard was applied before 1940 to other tree
species such as rubber and quinine, it was only in 1949 that the first pine seed orchards
were planted in Sweden. The use of this technique was intensified as from 1957 in the
south-eastern United States, in line with the expansion of pine plantations in response
to the establishment of many large paper factories in the region. The Cooperative
Programme of the University of North Carolina alone in association with 32 forest in-
dustry ventures in the region, has a total of 1 700 ha of pile seed orchards (185
orchards in three States) . This scale is justified by the seed needs of an annual
afforestation programme of 180 000 ha (Sprague ejt al , 1978).
The United States Forest Service figures for 1972 (Wright, 1976) list about 3 000
ha of Pinug taeda and . elliattii seed orchards that could produce seeds for at least
300 000 ha each year.
Howvr, the Queensland Department of Forestry in Australia can claim the honour
of having had the first fully productive seed orchard. The dona 1 orchard of Pinus
elliottii near Bttnrah has been producing at least 20 kg/ha/yr since 966. Thi. is the
**d used fey the D*parttnt for its plantations (Brown and Eldridga, 1977).
- 162 -
Another species that has been used intensively in afforestation programmes in
Australia and New Zealand is Pinus radiata. These countries have 625 ha of clonal or-
chards from grafts, producing more than 4000kg of seed per annum (Brown & Eldridge, (1977)
In 1975 second generation seed orchards began to be planted in the United States
and in Australia.
To date, almost all the information available concerns conifers, especially pines.
Forest hardwood seed orchards are recent and on a smaller scale; Tectona has been
grown in New Guinea, Thailand, Nigeria and India, and Gmelina in Nigeria. Mention should,
however, be made of work on eucalypts in Australia, South Africa, Morocco, Portugal and
more recently, Brazil, where private enterprises, with the collaboration and advice of
IPEF (Institute de Pesquisas e Estudios Florestais), are cooperating on a major effort
in this field.
World experience (despite the fact that this technique is relatively new) and the
results obtained with different species in different regions indicate enormous potential
for seed orchards for better yields of forest crops, and, therefore, more effective use
of sites and cheaper forest production.
PLANNING AND DESIGN
Generally speaking, the design of a seed orchard is based on the assumption that
all the clones or progeny composing the orchard will flower at the same time, be com-
pletely inter-fertile with all their neighbors, produce the same amount of pollen and
ovules, yield the same amount of seed, have the same degree of compatibility when
grafted, the same type of growth and a similar form of crown, etc.
Experience shows that this is seldom the case, since for each species and each
region, these characteristics must be studied in detail and all possible information
collected on clonal behaviour, compatibility, combining ability and any other information
inherent in the production of seed and its future behaviour, in order that the design
of the next and subsequent orchards can make maximum use of the available data.
Clones or seedlings
There is a wealth of literature on whether to use clones or seedlings from selected
trees, (Toda, 1964)., but in general it is agreed that only in special circumstances is
it desirable to use seedling progenies. Unless control pollinated seed is available, the
offspring will very likely be genetically inferior to the parents. Secondly, selections
in the orchard are applicable only to this environment. N ikies (1974) agrees that in
certain cases the seedling seed orchard has a special place in improvement programmes
when working with species that root with difficulty and that may flower and set fruit
early. The same author maintains it is a mistake to believe that the seedling seed
orchard is the low-cost solution for small breeding programmes, since great expertise
is essential to a successful orchard.
Many aspects of the establishment, design and management of seed orchards are
common to both clonal and seedling seed orchards. However, there has been more ex-
perience with the former, basically with grafting, although rooted cuttings have also
been used.
In practice, very few seedling seed orchards from selected trees have been estab-
lished. The decision is made not on genetic grounds but springs rather from a reluctance
to use a little-known technique where the investment is bound to be long-term.
Number of clones or seedlings
It is generally agreed that the right number of clones or families to obtain the
maximum genetic gain and avoid the bad effects of inbreeding, is between 15 and 20. There
are, however, good reasons for establishing a greater number of pheno-types to start
with. Possible incompatibility between the graft and the stock, flowering habits, and
also the need to thin clones, (leaving those with better general combining ability),
justify starting a seed orchard with a greater nunber of clones than is usual (in general
practice 60 to 100 clonesjl "~
- 163 -
According to Lindgren (1974), genetic gain is maximized by using a smaller number
of clones , provided they flower simultaneously and their orchard output is not used for
further selection. Five clones provide enough genetic variation to ensure adaptation to
different sites, resistance to diseases, changing uses for wood, and so forth.
An extreme illustration of this concept is the use of two clone seed orchards with
high specific combining ability for maximum genetic gain of known geno-types.
On the other hand, for progenie of selected trees, Shelbourne (1969) has advocated
starting the orchard with more than 200 families, with intensive intra- and inter-family
selection.
Initial planting distance
On the assumption that the best final spacing for most cultivated species is 10
metres, depending on expected intensity of selection, or earlier fruit-setting through
better pollination, the initial distance can vary from 2 to 6 metres. This spacing has
been used successfully for pines. As a general rule, it may be said that the space
between plants should be sufficient to allow free and complete crown development, with
maximum sunshine, to ensure a plentiful crop of fruit for several years before thinning
becomes necessary.
The minimum distance is often limited by the space needed for mechanical control
of weeds and grass, applying herbicides and fungicides, and seed collection.
Planting can be in squares, rectangles or quincunxes; Nikles (1974) mentions a
hexagonal distribution for better mechanical seed collection. However, after selection
and culling of undesirable genotypes, the final distribution is completely irregular.
Experimental design
The design most widely used throughout the world for clonal orchards and seedling
seed orchards is the randomised complete block. Shape and distribution within the 4 block
should be such as to ensure cross pollination in which all clones participate. To avoid
self ing, it is desirable to alter the random design where ramets of the same clone are
close together.
Some systematic distributions can be recommended within the block when the initial
number of clones is maintained until the end, otherwise the thinning of clones disrupts
distribution as much as random design.
The size of the orchard depends on the forester's need for seed; in other words,
it is in direct proportion to the projected afforestation annual programme.
The number of blocks will depend on the quantity of seed that experience of the
species and site dictate.
VEGETATIVE PROPAGATION
As indicated above, most orchards are established by using clone ramets. The tech-
nique commonly used is grafting, with various methods depending on the species and the
region.
Some species, mainly due to incompatibility between stock and graft, cannot be
successfully propagated by this method, and therefore other techniques must be used.
This was the case with some widely cultivated species such as Pinus radiata, Eucalyptus
S rand is and Pseudotsuga roenziesii, in different regions of the world. Sometimes, the
ecline and death of the clone take place over a fairly long period of time, complica-
ting the operation of the orchard and post-poning definitions. With Pinus radiata,
the problem was solved by making use of the ease with which its cuttings take root, and
the same method was followed with Eucalyptus grand! s, although seedlings from selected
trees have also been used to establish seed orchards. Considerable efforts have been
made with Douglas fir to improve grafting techniques, detect incipient incompatibility,
replace defective grafts and find compatible roots tocks. With this species, seed
orchards were often established from seedlings.
- 164 -
The methodology, techniques and problems of vegetative propagation have already
been discussed in another lecture.
ESTABLISHMENT
Location
In deciding on the location of a seed orchard, the following factors should be
borne in mind: climate and soil for high seed production; isolation from undesirable
pollen sources; facility of access and nearness to work areas.
The ecology of the site where the seed orchard of a particular species, variety
or provenance is planted, should be such as to encourage optimum development and maximum
fruit setting. Where the species are native and the orchard is located within the
range of the species, the best site can probably be identified quite easily. With the
more extensively cultivated species, there is already sufficient knowledge available to
reduce the danger of choosing a bad site. However, for species or provenances of recent
introduction and little history of. cultivation in other regions, choice of the seed
orchard site should have top priority.
As regards climate, it is usually unwise to select sites exposed to extreme tem-
peratures, or low areas exposed to frost. Some species, like Pinus elliottii and
tae< * a > d not appear to be affected by slight changes, whereas . caribaea is very
sensitive to frost.
Although one may generalize with regard to the best texture, quality and fertility
of soils for better fruit production, each species may have different requirements in
this respect.
Concerning the isolation factor, it is difficult to achieve perfect isolation
(which would be theoretically acceptable) and completely avoid contamination by un-
desirable pollen. The orchard would have to be situated outside the forest area or in
a remote or not easily accessible place. Experience shows that, in orchards with normal
flowering and fruiting, complete isolation can be sacrificed in favour of better siting
as regards accessibility and nearness to working places. The minimum distance accept-
able varies with species, topography, winds, etc. It has been found that for Pinus
caribaea var. hondurensis a distance of 200 metres is sufficient (Nikles 1974) to
achieve adequate isolation .
The site should be protected from wind, fire and animals, and near places of easy
access and intensive work such as nurseries. The site must be visited after for manage-
ment, fertilization, irrigation, pruning, collection of branches for propagation, polli-
nation, and regular inspection. Pest and disease control are also easier. Proximity to
labour force also facilitates seed collection, drying and extraction.
Site preparation
The site selected for the orchard should be prepared according to normal affores-
tation practices, with no skimping on preventive control of rodents, ants and other pests,
Clearing and weeding should receive the same care lavished on intensive agricultural
crops .
Site preparation should facilitate mechanical work as much as possible. Paths,
fire-breaks, etc., should be methodically laid out.
In clay or poorly-drained soils, it is desirable to chiselp lough the planting rows.
It is also desirable to use land with little or no slope; however, if no flat ground is
available, contour ploughing, or, in extreme cases, terraces, to check or prevent soil
erosion is recommended.
Planting
The primary objective should be maximum survival of the material planted. If this
material consists of nursery plants (seedlings of selected Plants), or grafts in pots
- 165 -
or clumps of soil, planting should be timed to coincide with optimum weather and soil
moisture conditions to ensure maximum success, as replanting in subsequent years has
proved to be ineffective. It is therefore desirable to plant a greater number of ramets
closer together.
On some occasions, good results have been obtained by planting the future root
stocks straightaway in the spot where they will be growing permanently and later grafting
III jiitu. This is particularly true for species that do not take well on transplanting.
In this case it is advisable to plant 2 or 3 roots tocks in each spot, as a precaution
against unsuccessful grafts or incompatibility, or in order to select the best graft, etc.
This latter method has produced good results in Queensland, with 85 to 100% success in
grafts of Pinus caribaea and Aj^aucaria cunninghamii , and also in the southeastern United
States with Pinus elliottii and . taeda. It has the advantage of ensuring more rapid
growth, earlier flowering and no root curling. (Nikles 1974). Among the disadvantages
are the need for more experienced grafters and more frequent visits to tend the grafts
in the field; also, weather conditions cannot be controlled (temperature, winds).
When dealing with little-known species, unpredictable climates or unskilled staff,
it is therefore preferable to use material which has been grafted, or has already taken
root, in shelters or nurseries.
MANAGEMENT AND HARVESTINO
Cu 1 1 i vat i on t e chn ique s
Plant cover
It is advisable to weed the plantation the first year, especially around the shoots
or seedlings, but experience has shown that later on it is better to let the grass grow
naturally (provided there are no aggressive rhizomes) or to sow grass, which can.be kept
short. This practice reduces or prevents soil erosion; affords excellent fire protection,
adds organic matter to the soil and keeps the orchard in excellent condition for work.
Weed killers or a mulch of straw and/or manure around the plants following super-
ficial raking - never deep raking or hoeing that could damage the roots - are recommended
when drought makes weeds over-competitive.
In areas with periodic droughts, or in dry years, it is preferable to irrigate the
plantation to conserve soil moisture rather than leave the ground bare (Nikles 1974).
Fertilization
Fertilization can be very effective in increasing seed production on skeleton or
poor soils. The type, rates and timing of fertilizer applications, vary according to
species and site. Experiments described by Nikles (1974) are summarized below.
In the region of origin of Pinus elliottii, the equivalent of 1 kg of nitrogen
(3 kg of ammonium nitrate) per tree is applied annually at the end of the spring or the
beginning of the summer. On phosphorus deficient soils, (i.e., less than 8 to 10 kg
of extractable P 2 05) and potassium (less than 60 kg of K 2 per hectare), 2,5 kg of
super phosphate and 1 to 1.5 kg/tree/yr of -potassium chlorate are applied. Lime is also
used where acidity is high or there is a deficiency of Ca and Mg.
In Queensland, an appreciable increase in seed production has been achieved through
NPK applications.
For Pi nu * taeda, the Cooperative Programme of the state University of North Carolina
has reached the conclusion that the best way to achieve good seed production is to ferti-
lize (NPK) in years of normal rainfall, and irrigate during periods of drought.
In Texas specific recommendations for fertilizer applications are made in the
light of results of leaf analyses. It is a common practice to apply nitrogen to promote
flowering.
- 166 -
Pruning and thinning
Neither pruning nor crown reduction nor girdling have produced the anticipated in-
crease in seed production of species with high apical dominance such as Pinus elliottii,
ZJL tae< * a an< * radiata. On the other hand, bending down TP. radiata branches has proved
successful in facilitating cone collection; as has pollarding which forces the tree to
produce cones lower down. Low branches are only pruned to facilitate access and move-
ment within the orchard.
Clonal selection in the orchard will depend fundamentally on the result of the
progeny tests, which will eliminate clones with poor general combining ability. The seed
production of each clone should also be borne in mind. If the clone produces not seeds
but male flowers (pines) and has good combining ability, it should not be culled. In
the selection of clones, the degree of incompatibility and other factors producing weak
or low-yield ramets should also be borne in mind.
Protection
Seed orchards are highly specialized, expensive and very important for afforesta-
tion programmes, so they are far more carefully protected than ordinary forest plantations,
This is not difficult, since the orchard is frequently visited by trained staff who can
easily spot any problem that could affect the normal development of the orchard. Damage
from diseases, insects, animals and birds, machine damage to trunks and branches and
wind or frost damage can thus be remedied or prevented.
Fires should be prevented or controlled, by keeping fire-breaks clear, not allow-
ing weeds to grow too tall, and having fire- fight ing teams always ready for emergencies.
Measures should also be taken to avoid erosion and excessive compaction of the soil.
Flowering, pollination and handling of pollen
The seed orchard manager needs a good understanding of the environmental factors
controlling flowering and seed set, the movement of pollen and the degree of inbreeding
and outcrossing. These are aspects that need to be studied in greater detail, and the
reader is therefore referred to Chapters 7, 8, 9 and 11 of Faulkner's book (1975).
Until now, pollination has seldom been considered sufficiently reliable for proper
seed setting to reduce self ing and avoid undesirable pollen. Brown and Eldridge (1977),
hold that it is possible to pollinate 1 000 flowers with 6 of pollen. One man with simple
equipment can pollinate 5 000 flowers daily. They also maintain that this technique,
with some refinements, will soon become part of seed orchard management, particularly
when the orchards are young.
Harvesting of fruit and seed
Harvesting techniques depend on species, climate, site, size of the orchard, local
or national economic conditions of seed. Mechanized harvesting is justified if labour
is scarce or is expensive, or if very fast harvesting is necessary because the seeds
ripen and are dispersed quickly in extensive orchards.
Some pines species have cones that fall very easily. These can be harvested by
hand, or with mechanical tree shakers. However, in some serotinous species the cones
are not easily detached from the tree. In these cases, elevating work platforms can
be used. The Australian experience with P. radiata indicates that 6 kg per day can be
harvested with this machine whereas only T.8 kg of seed/man/day can be collected by
hand (K. Willcocks in Brown and Eldride 1977). In the southeastern United States,
studies are being made on the use of a suction harvester to pick up the seeds of Pinus
taeda from the ground. Where harvesting time is limited because of rapid dehiscence
of the cones or fruits, they can be harvested green so long as the orchard staff are
familiar with off-tree fruit ripening techniques.
With eucalypts, that coppice well and produce a lot of seed, Brown and Eldridge
(1977) recommend felling the tree to facilitate harvesting. With serotinous cone
species that have very little fruit, and where labour is expensive, harvesting is at
two or three-year intervals to reduce costs.
- 167 -
Seed production
This varies with the species, age of the trees, clonal composition, site, management,
etc* It i* important for any organization managing an orchard to determine production
potential, so as to match orchard size to the forestry programme. Nikles (1974) reports
that a high-yield seed orchard of Pinus elliottii in Queensland produced 145 Ibs/acre
(approximately 160 kg/ha); a normal yield for this species in Australia is 75 Ibs/acre
(83 kg/ha). With Pinus caribaea var. hondurensis the experience of the Byfield orchard
indicates that between 13 and 34 Ibs/acre/yr can be obtained, with an average over 8 years
of 23 Ibs (25 kg/ha). Other younger orchards obtained lower seed productions. The
average yield of Pinus radiata, after 10 years, is 20 Ibs/acre (22 kg/ha). In the United
States, the average for Pinus^taeda, a higher yielding species than . elliottii is
estimated at 55 kg/ha/yr.
Seeds from orchards are usually bigger, and have better germination and viability
than seeds from natural forests or plantations.
Seed extraction and utilisation
When cones are collected from an expensive seed orchard, maximum care should be taken
in extraction and subsequent operations so as to maximize viability. The seed should be
extracted at low temperature under dry conditions. Seed should be stored in accordance
with the requirements of each species, with adequate protection from insects and rodents,
and proper disinfection.
Keeping records
Efficient seed orchard management and utilisation means keeping careful records,
including: charts of ramet location, listing cultivation techniques, fertilization, costs,
phenological observations, incompatibility, fruit production, relative seed production
from each clone, etc. Experience has shown that the more information recorded, the better
the management and future upkeep of the orchard.
BIBLIOGRAPHY
Brown, A.G. & K.G. Eldridge, 1977. Seed orchard design and management. International
Training Course in Forest Tree Breeding, Selected Reference Papers, Australian Develop-
ment Assistance Agency. Canberra, pp. 147-154.
Faulkner, R. 1975. Seed orchards. U.K. For. Comm. Bull. N. 54. London.
Lindgren, D. 1974. Aspects of suitable number of clones in a seed orchard. Proc. Joint
IUFRO Meeting 52.04-1 ,-2, -3, Stockholm.
Nikles, D.G. 1974. Seed orchards-concept, design, establishment and management. FAO/
DANIDA Training Course on Forest Tree Improvement, Kenya, pp. 176-195.
Shelbourne, C.J.A., 1969. Tree breeding methods. Tech. Paper no. 55, For. Res. Inst.
Rotorua, New Zealand.
Sprague, J., J.B. Jett and B. Zobel. 1979. The management of Southern pine seed orchards
to increase seed production. Proceedings of A Symposium on Flowering and seed develop-
ment in trees. Ed. Frank Bonner, US Forest Service, IUFRO, Mississipi State University,
Mayo 1978, pp. 145-192.
TodaR. 1964. A brief review and conclusions of the discussion on seed orchards. Silvae
Genetica 13 (1-2): 1-3.
Wright, J.W., 1976. Introduction to Forest Genetics. Academic Press Inc. N.Y. pp.463.
Zoble, B., J. Barber, C.L. Brown and T.O. Perry, 1958. Seed orchards, their concept and
management. J. For. 56: 815-825.
- 168 -
PROGENY TRIALS
M. Quijada R.
Institute de Silvicultura
Universidad de Los Andes
Merida, Venezuela
CONTENTS
Page
Definition and importance 168
Crossing value . 1 69
Estimation of genetic statistical parameters 169
Evaluation of groups or lines 169
Evaluation of individual trees 169
Types of progeny 1 69
Crossing systems 1 70
Experimental designs 1 70
Cultivation techniques 1 70
Evaluation 171
Records 171
Basic select bibliography 171
Annex I Specific models 1 72
1. Variance analysis for a factorial genetic design in a completely random and equal
replication statistical model.
2. Variance analysis for a modified diallel genetic design and a completely random
statistical model.
3. Variance analysis for a genetic design of polymixia or free pollination in a
completely random statistical model.
DEFINITION AND IMPORTANCE
The purpose of progeny tests is to estimate: (I) The genetic value of an individual
on the basis of the behaviour of its offspring, or (2) the genetic value of individual
half-siblings or full-siblings.
In programmes based on phenotypic selection, where environmental influence is un-
known and where selection is consequently not very reliable, progeny tests are indispens-
able.
- 169 -
Progeny tests facilitate:
(a) Evaluation of the crossing value (combining ability) of trees;
(b) Estimation of genetic statistical parameters (variance, correlation, etc.).
(c) Evaluation of sub-populations or lines for specific uses in plantations;
(d) Evaluation of individuals for the purpose of continuous selection.
Crossing value
This is the value of a tree, determined by the average of its progeny obtained from
crosses with one or several other trees, following a specific plan. The crossing value of
individuals is determined by specific crossing systems (genetic designs). From this, we
can determine the combining value of an individual which is a statistical parameter that
indicates the ability of an individual to transmit certain of its characteristics to
progeny resulting from crosses with one or more other individuals. This information is
used as a basis for selective thinnings in seed orchards.
Estimation of genetic statistical parameters
This is done through a series of theoretical assumptions on the components of the
phenotype and constitutes the mathematical basis of genetics. It involves the use of
special types of genetic and experimental designs.
These require determination of the components of variance and co-variance, using
Expected Mean Squares in Variance and Co-variants Analysis Tables.
The determination of such parameters as heritability (h2) , makes possible eventual
estimation of genetic gains or the response to breeding schemes as regards both how many
in numerical value and how much improved.
Evaluation of groups or lines
In specific cases, the results of particular combinations between clones or families
resulting in progeny outstanding for a specific characteristic (high value of specific
combining) can be used.
Evaluation of individual trees
For the purposes of continuous selection, individual evaluation is made as and when
advanced seed orchards are established, i.e. in future generations.
This requires continuous selection within progenies of the most outstanding individ-
uals. It can lead to problems of inbreeding in orchards of an older generation, but
this can be counteracted by having a sufficiently broad basis for initial selection, and
separating the seed production population from the breeding population (see paper entitled
"Planning and Strategies of a Forest Tree Improvement Programme 11 ).
TYPES OF PROGENY
The quantity of information that can be provided by trials depends on the type of
progeny, among other things.
We may distinguish two principal types: semi-fraternal (half-sibs), and fraternal
(full sibs).
Half-sibs are progeny with a known parent in common (the mother tree). They are
typically freely pollinating, the exact degree of inbreeding is unknown.
Information obtained will depend on the value of the seed tree.
I/ For more detailed information, see paper entitled "Quantitative Genetics",
- 170 -
Full sibs are progeny with two known parents in common. This can only be achieved
through controlled pollination.
In this case, the tests will produce more information, as it is based on the values
of both parents. However, most progeny tests made to date have been with half-sibs, since
the establishment of full-sib tests involves considerable work and expenditure.
CROSSING SYSTEMS
The way in which parent trees are combined influences the results of tests, since
it determines the type of progeny obtained.
There are two important general types of crosses in the field of forestry: (1)
crossing systems with an unknown father, and (2) crossing systems with a known father.
In the first case the principal systems are: (a) free pollination and (b) poly-
crossing; in the second (c) diallel designs (complete, modified, partial diallels),
(d) factorial designs, and (e) mating of a single pair. For detailed information on
these designs, see the paper entitled "Systems and Designs of Controlled Crossing".
EXPERIMENTAL DESIGNS ~
The environmental component in progeny tests is controlled by the use of experi-
mental designs which seek to reduce the non-genetic effects. Randomization (random distri-
bution), replication (repetition) and local control (blocks) should be taken into con-
sideration.
The use of blocks facilitates control of local variability, exposing the progeny
to fairly standard site conditions. Ideally, replication makes it possible to expose
individual progeny to different site conditions, which occur quite frequently in -the same
plantation. The genetic and environmental components of the phenotype can thus be
estimated. Randomization is indispensable for obtaining estimates not vitiated by ex-
perimental error.
The testing site should be representative of the area where the seed will eventually
be used in a normal plantation. The best experimental design usually depends on the
number of progeny, plot size, and site variability. The complete blocks design is often
used for low numbers of progeny and trees per progeny, and sites of average to low varia-
bility. For a large number of progeny (more than 25), large plots and very heterogeneous
sites, it is advisable to use incomplete block designs such as lattices.
Plot size can range from one tree per plot upwards. Low numbers, particularly in
row plots, are desirable if the value of intra-progeny competition is to be ascertained.
High numbers are desirable with square or rectangular plots in estimating genetic statis-
tical parameters where the value of competition is a bias factor. In this case, one con-
dition is a variable number of trees in the centre of each plot of a given progeny, that
are not in direct contact with other progeny, or else are surrounded by buffer areas of
trees of their own kind.
In order to cover the maximum intra-progeny variability, it is recommended that the
number of replications be increased for smaller plots.
Replication can be spatial (control of local environmental variability) and temporal
(control of climatic or biotic variations).
CULTIVATION TECHNIQUES
Spacing should be the usual planting distance in a normal programme. Tending
should, as far as possible, be the same as for large-scale plantations. "Luxury tending"
should be avoided.
Protection against fire, wild animals, etc., should be effective.
\J For more detailed information on experimental designs, see the paper on this subject.
- 171 -
EVALUATION
The characters evaluated will depend on the actual and potential uses of the species
under consideration. There are, however, several (general) characters common to the
majority of uses, such as survival, growth and vulnerability to pests and diseases; other
(specific) characters will depend on the use envisaged.
Evaluation in the field usually begins six months or a year after the test, with
mortality and attack counts, and measurement of initial growth. Evaluation is usually
repeated in the second and in the fifth, seventh and tenth years, including other charac-
teristics in the measurement, as the individual trees develop. For most characters, the
minimum age for making reliable evaluations is considered as equal to one-third of the
final rotation of the species. For some characteristics, such as wood quality, it is
usually necessary to wait until the end of the rotation, with periodic evaluations every
three or five years.
Evaluations are also made in the nursery, for correlative comparisons with field
behaviour. These evaluations can also be used for selecting within the nursery itself.
Even more important, these studies can include seed characteristics such as weight and
size.
RECORDS
A detailed record of the material and the procedures used is indispensable, a file
should be kept for every test made. This file should include a summary of the objectives
of the test, and all observations made during the course of the test, data and analysis
of evaluations, etc. This procedure will make it easier to come to the right conclusions
at the end of the test.
BASIC SELECT BIBLIOGRAPHY
Becker, Walter A., 1967. Manual of Procedures in quantitative genetics. Washington State
University Press, Pullman, Wash., USA. 130 pp.
Cockerhan, C. Clark, 1963. Estimation of Genetics variances. Statistical Genetics and
Plant Breeding. National Academy of Sciences, Publication 982, USA. pp. 53-94.
Falconer, D.S., J960. Introduction to Quantitative Genetics, Ronald Press, 1971 New York,
365 pp.
Franklin, E.C., 1969. Quantitative inheritance, Heritability and Combining Ability.
Forest Tree Improvement Training Course, FAO-N.C. State University, Raleigh, N.C.
USA. Volume I: 96-107.
Mather, K. and Jinks, J.G. 1971, Biometrical Genetics Chapman & Hall, N.Y.
Matheson, A.C., 1977. Progeny Tests. Training Course in Forest Tree Improvement,
Australian Development Assistance Agency, Canberra, Australia; pp. 113-119.
Namkoong, G., E.B. Snyder, and R.W. Stonecypher, 1966. Heritability and gain concepts
for evaluating breeding systems such as seedling orchards, Silva Genet ica 15: 76-84.
Roberds, J.H., 1969.' Progeny Testing in Forest Tree breeding. Forest Tree Improvement
Training Course, FAO-N.C. State University, Raleigh, N.C. USA Volume I, 123-135.
Wright, J,W., 1976. Introduction to Forest Genetics. Academic Press, Incl. New York.
463 pp.
- 172 -
ANNEX I
Specific models
Variance analysis for a factorial emetic design in a
completely random and equal replication statistical model
Source of
variation
Degrees of
freedom df
Mean square
expectation E(MS)
Genetic
interpretation
Male
m - 1
<T e +rmh+rhm
2
o* m - Cov(fm)
Female
h - 1
222
(T e + r mh +rmh
6 h 2 - Cov(fh)- 1 VA
Male x Female
(m-1) (h-1)
<T e + r m h
CT m hVarCf)-Covfm
Error
mh (r-1)
*e 2
- Cov f h
7
Total mhr- 1
02-
Covfop"
Var,
D-
Component of variance for the source of variation indicated by the letter (s)
below.
Co-variance of half-sibs with a male parent in common.
Co-variance of half-sibs with a female parent in common.
Sibling variance.
Additive component of genotype.
Non-additive component of genotype.
- 173 -
2.
Variance analysis for a modified diallel genetic design
in a completely random statistical model
Source of
variation
Degrees of
freedom
df
Mean square
expectation
E(MS)
Genetic
interpretation
General combining
value
p - 1
ae 2 + r e 2 + r (p-2) g 2
2
ag -Cov f
Specific combining
value
Experimental error
p(p-3) 12 \ 0e 2 + r e 2
p(p-l) (r-l)/2 oe 2
ere 2 Var f-2Cov f
I'D
p-
02-
Cov f-
Var f-
VA-
V
3.
Number of parents.
Components of variance for the source of variation as indicated.
Half-sib co-variance.
Full-sib variance.
Additive component of genotype.
Genotype component.
Analysis of variants for a genetic design of polymixia or
free pollination in a completely random statistical model
Source of
variation
Degrees of
freedom
df
Mean square
expectation
E(MS)
Genetic
interpretation
Progeny
(P - 1)
2 2
cre z + r p
2
a p Cov fl
4 VA
Experimental error
f(r - 1)
oe 2
* Component of variance for the source of variance indicated,
Cov f Co-variance of half-sibs.
VA* Additive component of genotype.
- J74 -
GENOTYPE-ENVIRONMENT INTERACTION
M. Quijada R.
Institute de Silvicultura
Universidad de Los Andes
Merida, Venezuela
TABLE OF CONTENTS
Existence and importance
Determination
Rank position
Variance analysis . . .
Regression analysis
Bibliography
Page
174
176
176
176
177
177
EXISTENCE AND IMPORTANCE
Genotype-environment interaction may be defined as lack of uniformity in the res-
ponse of two or more groups of plants grown in two or more environments; one group may
show the best growth in one environment, but be mediocre in another.
The interaction may show itself in a change in the relative position of the groups
in each environment , or in differences in the degree of superiority, even when the rank
positions are equal in each environment.
In graph form, an interaction is represented by lines which cross or approach each
other, as the environment varies, following a given gradient.
Value of character tested
Value of character tested
1 2
Environments
1 2
Environments
Since it is more apparent, the first makes it necessary to select a distinct group
for each environment, in order to obtain the best gain in productivity.
The second is important as a guide for the extrapolation of data to environments
different from those tested.
The presence of genotype-environment interaction reflects the existence of environ-
mental variations, even over relatively small distances, and of variations in the require-
ments of different genotypes in both macro and micro environments. The same reasoning
may be applied to seasonal variations.
- 175 -
This is the fundamental reason why experiments need to be repeated both in different
areas (control of local environmental variation) and at different seasons (control of
cyclical variation).
Different levels of genotype-environment interaction may be distinguished in forestry,
due mainly to the genotypic constitution in question. Thus we note the existence of
species x environment, provenance x environment snf progeny x environment effects.
In terms of species, the differences with regard to adaptability to environments
are greater, i.e. there are species which possess a wide range of adaptation, so that the
detection of interaction necessitates tests over a wider range of environmental conditions.
If the genetic spectrum, i.e. the provenances and progenies, is reduced, interaction tests
become more sensitive, in that the ranges of tolerance narrow, and this sensitivity shows
up in different reactions to minor variations in the environment.
Of the environmental factors that indicate interactions, the most tangible are soil
conditions, since these can be detected within a plot, over relatively small areas.
Climatic conditions have to be more extreme in order to be detected, e.g. variations due
to extreme conditions of temperature (low or high), humidity (drought or flooding), etc.
The interaction may be considered as an indicator of the relative stability of a
genotype. If the interaction is close to zero, the genotypes are sufficiently stable for
the characteristics under consideration, i.e. their relative positions and the differences
in magnitude of responses are similar in the different environments; the only thing that
changes is the actual magnitude of the responses, which will depend on the particular
genotypes (species, provenances, individuals) and sites under consideration.
Genotypes with a wide range of adaptability, as shown by zero or insignificant
interaction, are called plastic, while those with a high susceptibility to marked varia-
tion in behaviour with slight variations in the environment are called rigid or strict.
Since the components of genotype and environment also play a part, though small,
in the response of individuals (evaluated in the pheno types), the interaction will never
represent 100% of the variation observed; however, the nearer one comes to this value,
the lower will be the stability of the genotype in different environments, and from the
point of view of the breeder, the less reliable it will be for large-scale propagation
under normal plantation conditions. It will be more suitable for plantations in limiting
environmental conditions, since these are established in relatively small areas and under
more honogeneous site conditions.
The present tendency of breeders is to produce sub-populations with a broad spect-
rum of adaptability, which, in the long run, can represent greater gains for lower costs.
If the interaction is not significant, selection will be made on the basis of the
mean of the responses of the genotypes in all the environments. This same criterion may
be adopted when the interaction is significant, but at levels very close to the lowest
critical limit of probability (usually 5%), in which case it may be ignored in practice.
The practical problems of using interaction as a yardstick are highlighted by
certain factors, the difficulty being to interpret it correctly. Very often the existence
existence of interactions is reported on the basis of a very few tests, located very
close to each other, which in practice would be of very dubious applicability from the
point of view of costs. In these cases it is recommended that the interaction be con-
sidered as part of the experimental error and the average yields from the sites tested be
taken as the basis for work*
Effective practical use of interaction requires essentially testing of the same
genotypes in a range of environmental conditions which represent marked variations, with
areas sufficiently large to be used for plantations.
- 176 -
DETERMINATION
Rank position
A simple, preliminary way of evaluating the presence of genotype-environment inter-
action is by using the rank position of each of various genotypes in each of various
sites or environments.
The genotypes are graded from best to worst, according to their yield in each site,
assigning them a number from 1 to n (according to the number of genotypes).
If there are noticeable changes in position from one site to another, this indicates
interaction.
Environments
Genotypes I
2
3
A 1
B 5
C 4
D 3
E 2
3
1
2
4
5
3
2
5
4
J
A disadvantage of this method is that rank classification gives us only n overall
idea of the magnitude of the changes from one environment to another, since the changes
will always be in constant units. In addition, there is no statistical information avail-
able on the significance of these changes.
A statistical approximation can be obtained by using rank correlation, using pairs
of stations. These may be all the pairs possible according to the number of environments,
or the environmental extremes in question.
Variance analysis
The importance and magnitude of the interaction are evaluated by using different
methods of variance analysis. One method is to combine variance analysis with a specific
experimental design (e.g. random complete blocks).
Sources of variation
Sites
(L-J)
Blocks/sites
L(r-J)
Genotypes
(t-1)
Genotypes x
sites
(t-l)(L-l) MSI MSI
2-2
rfe+Kl 61
est. MS
Errors L(t-l)(r-l) est .MS
Total Lrt-1 6F 2
The significance test gives us information on the overall importance of the inter-
action. We obtain its magnitude (0I 2 ) and effective participation in the total variation
(6F) by estimating the variance components, using the means squares ( E(MS) ) anticipated
from the variance analysis table.
- 177 -
Regression analysis
Another way of using the principle of variance analysis is through the regression
of the genetic values in the environments. For this it is necessary to establish the
same genetic entities (normally progeny) in various well-differentiated environments
(preferably more than two) . The individual values of a given characteristic of each
genetic entity (genetic values) are related to the overall means of each site (environ-
mental values) adjusting simple lineal regressions. This results in various straight
lines (as many as there are progeny), which are then compared with each other, analytically
or graphically. The graphic form is the most usual, and the presence or absence of inter-
action is indicated by the arrangement of the lines. If they are more or less parallel,
the genetic entities are said to be relatively stable (no interaction). If the lines
cross or markedly diverge from the parallel position, this indicates the presence of
interaction.
Analytical methods involve comparison of the lines to see whether they are to be
considered as having the same slope, and involve, for each genetic entity quantifying the
importance of the regression or deviation.
The most stable genetic entities are those whose regression approaches the unit and
the deviations zero. If these stabilities are associated with acceptable productivity
(best or nearly best), then they will have preference over others which, although the best
in a given environment, are unstable as a whole.
BIBLIOGRAPHY
Allard, R.W., 1964. Principles of Plant Breeding, John Wiley & Sons, Inc. New York. 485 pp.
Bur ley, J. & Wood, P.J. (Eds), 1979. Manual on species and provenance research, with
special reference to the tropics. CFI Tropical Forestry Papers No. 10 and 10A, Oxford,
pp. 127-130.
Falconer, D.S., 1960. Introduction to Quantitative Genetics. The Ronald Press Company,
New York. 365pp.
Goddard, R.E., 1977. GE interaction in slash pine. Third World Consultation on Forest
Tree Breeding, Canberra, Australia, pp. 761-772.
Hill, J., 1979. Genotype-environment Interaction - A Challenge for Plant Breeding. J.
Agri. Sci. Ji5: 477-493.
Namkoong, G., 1978. Choosing strategies for the future. Unasylva 30. (1 19/120) , pp. 38-41 .
Rink, George. Variation in open-pollinated progeny plantations of Virginia Pine (Pinus
virginiana Mill) . PhD Thesis, University of Tennessee, Knoxville, Tenn. U.S.A. 143 PP
W75.
Participants visiting CVG Pinus caribaea plantations in Uverito
- 178 -
BREEDING FOR DISEASE RESISTANCE I/
C. Palmbcrg
Forest Resources Division
Forestry Department
FAO
CONTENTS
Introduction
Variability of the pathogen
Variability of the host
Principles of selection and breeding for disease resistance
Assessing the control alternatives
Developing methods for evaluation of resistance 181
Selecting for disease resistance 181
Breeding for disease resistance 1 82
Maintaining resistance 183
Practical implications . 183
References 1 85
INTRODUCTION
A disease involves harmful physiological changes in a plant. The changes may be
caused by non-pathogenic agents such as adverse climatic or soil conditions, or by a
pathogen (Tarr 1972). The most important biotic diseases of forest trees are caused by
fungi (Cowling 1969). Research on the genetics of insect resistance and on resistance to
bacteria and viruses has not been nearly so extensive as on the genetics of resistance to
fungal diseases. Nevertheless, enough is known to indicate that the inheritance of
resistance to these agents differs in no major way from inheritance of resistance to fungi
(Allard 1964; Painter 1966). Major stress in this paper will be placed on infectious
diseases caused by fungi. The term 'breeding* will be treated in the broad sense, thus
including procedures of selection.
The general objective of a program to produce pest resistant trees is to reduce
losses to a pest by genetic manipulation of the host trees (Callaham e ail. 1966). The
probabilities of success in a breeding program will, to a large degree, depend on the
establishment from the outset of a clearly defined set of objectives and a realistic
order of priorities. The prospects for significant improvement of disease resistance
are better in forest trees than in most cultivated crops, because forest tree species
generally represent natural cross-pollinated populations which provide a wide range of
genotypes for basic selection (Schreiner 1966). Concurrently with improvement in disease
resistance, it is necessary to provide for improvement in yield and quality of the crop,
A realistic weight must be given to the importance of each factor in the breeding program
(Borlaug 1966).
Selection and breeding for disease resistance does not vary fundamentally from
breeding for any other character (Allard 1964; Dyson 1974). Consequently, any of the
various methods of breeding appropriate for the crop or species in question can be used
in developing disease-resistent varieties or population, once resistance-conferring genes
have been found (Allard 1964). However, it must be kept in mind that disease
jl/ Based on a paper presented at the Training Course in Forest Tree Breeding held in
~ Canberra, Australia, 1977.
- 179 -
resistance involves two genetic systems, that of the host and that of the pathogen, each
a result of the two basic variables of genotype and environment. The phenotypic ex*
pression of a disease will thus depend on the interactions between the host, the patho-
gen and the environment (Mclntosh 1971). As the pathogen populations constantly react
to changes introduced into the host populations (Dinoor 1975), resistance breeding is a
dynamic rather than a static process. A full understanding of the pest complex, including
the variability and genetic potential of the host, the biology of the pathogen and the
interactions between host, pathogen and environment, is a necessary basis for a program
on breeding for resistance.
VARIABILITY OF THE PATHOGEN
Diseases in forest trees can be grouped into two categories, those caused by
obligate and those caused by facultative parasites. Obligate parasites are restricted
to living tissue, whereas facultative parasites are able to colonise both living and
dead organic matter (Tarr 1972). The rust fungi are examples of obligate parasites in
forest trees (e.g. blister rust of pines and leaf rust of poplars), while facultative
parasites, which often are limited to certain provenances, include blight diseases (e.g.
needle cast disease of pines and stem canker of poplars) (Bjorkman 1966). The suscepti-
bility of trees to facultative parasites is generally greatly affected by their
physiological condition and growth vigour, and thus indirectly by the environment
(Schreiner 1966).
Pathogenic organisms have enormous potential for developing new, virulent forms
or races. Plant breeders must therefore be prepared to face new races of pathogen to
which their formerly resistant varieties or populations are susceptible (Allard 1964).
Once a fungal race becomes established, its prevalence is determined by the
varieties of the host plants being grown. An extreme case is the growing of a crop con-
sisting of inbred lines or one single clone in which the individuals are genetically
identical, possessing identical genes for resistance against prevailing races of the
pathogen. A host population like this will exert a strong selection pressure, and is
likely to give rise to highly specialised physiologic races of the pathogen which are
able to overcome the resistance of the host (Borlaug 1966).
The key to understanding variability in fungi, including variability in pathogenic
capabilities, lies in their reproductive systems. In addition to meiotic recombination
of the genetic material, variation may be added to the population by non-meiotic recom-
bination, by heterocayosis, in which two or more nuclei occupy the same cell, by cyto-
plasmic inheritance, and by mutation which, itself, provides the initial genetic
differences that are brought together in the various recombination processes (Buxton 1961).
The majority of the species are haploid organisms in which nuclear fusions occur to give
rise to a relatively short-lived diploid stage, but life-cycles ranging from completely
haploid to completely diploid are also encountered (Allard 1964).
VARIABILITY OF THE HOST
Variability of the host leading to resistance can be either active or passive.
Passive resistance is accomplished by structural barriers of a morphological or anatomical
nature. Active resistance is due to vital processes initiated in the host as a reaction
against a pathogenic agent, and is directed against the agent itself (G&uman 1950).
Schrefner (1966) lists the following host characteristics as the most common ones
contributing to disease resistance in forest trees:
(1) Variation of the host in the uptake of minerals under similar field conditions;
(2) Differences in water content of the host tissue;
(3) The presence of metabolic products in the host tissue other than phytotoxins;
(4) The presence of fungistatic or phytotoxic substances in the host;
- 180 -
(5) The production in the host tissue of phytotoxins;
(6) Pre-existence of formation of barriers that restrict the movement of the
pathogen into and within the host tissue;
(7) Effect of osroosed substances on germ tube growth and penetration of stomata;
(8) Escape mechanisms such as drooping needles or densely pubescent leaf surfaces.
The genetic nature of host resistance can either be vertical (specific resistance)
or horizontal (non-specific resistance, also called field resistance or tolerance)
(van der Plank 1963). Vertical resistance acts against specific races or varieties of the
pathogen; it is generally characterised by a close correspondence between the genes of the
pathogen and those of the host, and is often controlled by single genes, although many genes
may be involved. The genetic components that contribute to vertical resistance can
generally be isolated and studied in appropriate crosses (Schreiner 1966; Watson 1971).
Horizontal resistance is generally polygenic and many factors, such as control of
rate of growth of the hyphae through the host tissue and penetration through the tissue,
may simultaneously control resistance. Horizontal resistance generally gives a certain
amount of protection against a whole range of races of the pathogen (Watson 1971). No
direct correspondence exists between host and fungal genes, and although generally more
efficient against fungi and favoured both in agriculture and forestry, polygenic systems
are more difficult to utilise in breeding than the major genes controlling vertical
resistance (Heybroek 1969; Luig 1971; Watson 1971).
Present evidence for forest trees indicates that resistance may be due to a small
nunber of major genes (e.g. in blight resistance of poplar hybrids; see Heimburger 1966),
monogenic resistance (e.g. in resistance of western red cedar to leaf blight; see Soegaard
1966), polygenic resistance (e.g. in resistance of western white pine to blister rust;
see Bingham t al^. 1960), and cytoplasmic inheritance (e.g. in resistance of larch to
needle blight; see Lagner 1952). Both major genes and polygenes may simultaneously con-
tribute to resistance (e.g. in resistance of eastern white pine to blister rust; see
Heimburgher 1962).
While it has been generally accepted that different genetic systems operate in the
case of vertical and horizontal resistance to a single disease, the same genes may be
involved in both types of resistance (Watson 1966). A diversity of resistance genes in
the population will reduce the chances of the pathogen overcoming existing resistance by
a change in virulence or pathogenicity (Borlaug 1966; Watson 1971; Dinoor 1975).
In addition to being genetically controlled, horizontal resistance may arise from
imperfect synchronisation of pathogen and host, as in diseases which appear late in the
life of the plant or late in the growing season (e.g. poplar leaf rust). This verges on
disease escape, in which vigorous individuals may be able to replace diseased leaves or
other parts, and so produce a reasonable yield in spite of infection. Plants may also
escape infection if their susceptible phase happens to come at a time when the pathogen
is absent, present in insufficient amounts, or when environmental conditions are un-
favourable to it (Tarr J972).
PRINCIPLES OF BREEDING FOR DISEASE RESISTANCE
Assessing the control alternatives
Before embarking on a program to breed pest-resistant trees, alternative methods
of disease control should be carefully assessed. These include: (1) exclusion,
(2) avoidance, (3) eradication, and (4) protection through chemical or biological control
(McNabb 1964; Gibson 1975).
Exclusion through quarantine, i.e. restricting the introduction of plants, parts
of plants or soil into the country (as e.g. in Australia) may be effective against
diseases not yet established, provided there are adequate means of enforcing the legis-
lation. Exclusion through inspection and treatment of planting material and possibly
tools and machinery before transferring them from one region to another pre-suppotes
good information on the spread and life-cycle of the pathogen (Gibson 1975).
- 181 -
Disease avoidance through a change of species is usually effective if an accept-
able alternative species is available (e.g. the change from one species of cypress to
another in East Africa discussed later in this paper; the change from spruce to pine or
larch on some sites infested with root rot in Northern Europe; see Palmberg 1969).
Adjustment of e.g. sowing and planting times to reduce the risk of infection can some-
times be successful, as can a change in management practices (e.g. a change in pruning
routines of cypress in East Africa; see Rudd-Jones 1954).
Eradication of other than nursery diseases must be achieved at a very early stage
of the disease outbreak, and calls for an exceptionally efficient survey organization to
be effective, e.g. eradication of diseased individuals of elm to stop the spread of dutch
elm disease have largely failed (Neely 1975). Eradication of the secondary host in rust
diseases has, however, proved fairly effective (e.g. removing aspen in Northern Europe and
Ribes spp. in USA from areas close to pine plantations to avoid damage by pine twist and
fusiform rusts; Day 1972).
Protection by the application of fungicides is generally regarded as too costly
outside the nursery. However, in diseases which attack the trees during a limited period
of their life this approach can sometimes be economically acceptable (e.g. spraying of
radiata pine against needle blight in New Zealand; see Gilmour and Vanner 1971).
Protection by the application of insecticides is usually considered feasible during out-
breaks of insect epidemics (Benedict 1964). Biological control has been very success-
fully employed against insects (Franz 1964) (e.g. control of the Si rex wasp in Australia;
see Anon. 1974) .
The above control methods are in many cases practiced parallel with programs on
breeding for disease resistance, and some of them provide only short-term solutions to
a problem. Generally the most economical and, in many respects, ideal method of
achieving disease control is the development of resistant biotypes of the host (Gibson
1975).
Developing methods for evaluation of resistance
Once a decision to breed for resistance has been reached, the development of rapid
and accurate methods for evaluating resistance in the host plant is the most important
immediate problem. Resistance is not an intrinsic character, but is subject to environ-
mental influences on the host, the pathogen and the host/pathogen relationship (Schreiner
1966). In addition to basic research on the establishment and progress of parasitism,
the development of such evaluation methods will therefore require knowledge of the effects
of environmental factors on host resistance and of the pathogenicity and virulence of
the parasite (Callaham t ad. 1966; Schreiner 1966).
As expression of resistance is phenotypic, the same expression may be the outcome
of interaction between many different host and pathogen genotypes (Dinoor 1975). The
identification of genes for resistance therefore involves crosses and genetic analyses,
which in turn require the development of efficient techniques for selecting, propagating,
breeding and progeny testing the host (Schreiner 1966).
Identifiable resistance genes depend on the pathogenic strains present; in the
absence of the pathogen, resistant genotypes are indistinguishable from non-resistant
(Allard 1964). Programs for identifying resistance should therefore always be based on
artificial inoculation (Allard 1964; Dyson 1974). As immunity to a disease is usually
an unrealistic objective, it is necessary to set up minimum practical standards of
resistance for each host/pathogen relationship. For both economic and biological reasons
these standards must be kept flexible (Schreiner 1966).
Selecting for disease resistance
Aft^r developing methods for recognising and identifying disease resistance, the
next step is to locate resistant material in existing populations of the host (Allard
1964).
Resistance to disease in plants is more common than susceptibility, most wild
plants being resistant to ttost diseases (Cowling 1969). Our most disastrous forest tree
diseases have resulted from the introduction of foreign pathogens into populations which
carry no effective genes for resistance to this pathogen (e.g. chestnut blight, dutch elm
disease, white pine blister rust) (Borlaug 1966). Widespread planting of fast-growing exotics
- 182 -
sometimes under unfavourable ecological conditions and of populations possessing a very
narrow genetic base, also easily leads to phytopathological problems (e.g. needle blight
in radiata pine in East Africa, poplar leaf rust in Australia) (Bjdrkman 1966).
Genes- do not occur at random in populations (Qualset J975). Long association
between pathogen and host iff likely to lead to the development of mutual tolerance , with
consequent elimination of highly susceptible host genotypes through natural selection
(Tarr J972). Genes for resistance to pathogens and pests are therefore likely to be
found in highest frequency in regions where the plant in question has been grown on a
large scale in the presence of the pathogen, or by an extensive search of natural
populations.
A large number of individuals should be screened at high levels of artificial
inoculation to augment any search for resistant individuals in the forest. Resistance
teats must be designed in consideration of the biology of the host and the biology of
the pest; for example, the age of the host at which infection and diagnosis are to be
made must be known. The trials should be replicated to sample a range of environments.
The design must allow for statistical procedures to deal with variation in host, pest
and environment, and their interactions. Tests on clone x environment interactions
should not be considered conclusive until 1/3 to 1/2 of the estimated rotation elapses
(Callaham et al. 1966).. Screening for vertical resistance should be undertaken in each
planting region separately, whereas screening for horizontal resistance often can be
centralised. However, factors like climate and the amount of virulent inoculum present
in various planting regions may alter the level of resistance found at a central testing
site (Dinoor 1975).
If vegetative propagation is feasible for mass -product ion of planting material
(as e.g. in poplars) and if the resistant trees have desirable phenotypes, they can be
used immediately to develop pest resistant test populations. However, in the case of
species reproduced by seed, it should be remembered that trees not only vary in resist-
ance to a pathogen, but also in their ability to transmit resistance to their progeny
(Wood 1966). It is therefore essential that phenotypically selected trees of these
species are progeny tested. If resistant progenies emerge, their parents can tentatively
be designated 'resistance transmitters* (Schreiner 1966), and after further tests pro-
pagated into orchards to produce resistant Fj seed (Callaham e aU J966). Trees ex-
hibiting high general combining ability for resistance will be of greatest use. Specific
combining ability can be utilised only through vegetative propagation or through controlled
crosses.
Breeding for disease resistance
Variation in the degree of resistance has been attributed to the kind and number
of resistance genes present in the parent (Wood 1966). Breeding can be used to ensure
that cultivars possess maximum genetic diversity for resistance genes (Watson 1966).
In polygenically controlled systems a selection plateau may be reached beyond
which further advances are insignificant. This plateau may or may not confer a degree
of resistance adequate for practical purposes. Introducing new variation through hybridi-
zation of the most resistant individuals and additional populations of the host which
show some degree of resistance may overcome this plateau (Heimburger 1962).
If sufficient disease resistance cannot be found in existing populations of the
host species, genes for resistance can sometimes be introduced through inter-specific
hybridization. If the resistant material thus developed is not commercially acceptable,
two procedures are available: the Fj hybrids can be crossed to produce an F generation
where segregation and recombination may result in acceptable resistant phenotypes, or
the Fj hybrid can be backcrossed to the desired parent species (Allard 1964). Renewed
screening for disease resistance must be undertaken in these new populations.
Mass-producing inter-specific Fj seed possessing both desirable phenotypic charac-
teristics and disease resistance may be done either through controlled pollinations or,
where the two species involved flower simultaneously, they may be interplanted in
orchards to produce hybrid seed spontaneously. In some cases single clones of one species
can be interplanted within stands of the other species; this procedure facilitates roguing
of non-hybrid seedlings in the nursery (Callaham e al. 1966).
- 183 -
Both intra- and inter-specific hybridization nay be used to transfer genes for
resistance to desirable phenotypes.
Maintaining resistance in improved populations
There is evidence of sufficient variation in resistance between species, races
and individuals to warrant the use of resistance breeding to combat practically all
important plant diseases (Schreiner 1966). Whenever satisfactory disease resistant
varieties are available they have been preferred over other means of control, because
once developed they add little or nothing to the cost of production (Allard 1964). In
crop plants which to a great extent consist of genetically homogeneous planting material
and inbred lines, resistance breeding demands continuous study and mobilisation of new
and different sources of resistance (Dinoor 1975). For example, the average maximum
duration of effective protection by any given type of stem rust resistance in wheat in
USA and Canada has been about 15 years; the situation has been even worse in respect of
areas of the sub-tropics and the tropics, where the disease escape mechanism does not
function effectively and where the gene pools for pathogenicity are broad and persist
from year to year (Borlaug 1966).
As long rotations are the rule in forestry and effects of any long-term climatic
changes on the host, the pathogen and their inter-relationships may be pronounced, the
question of maintaining resistance is of vital importance (Schreiner 1966; Painter 1966).
The probability of attaining and maintaining resistance by breeding for extended
periods of time is generally directly proportional to the diversity of germplasm avail-
able to the breeder (Painter 1966). Whereas the basic unit in crop plant breeding is an
inbred line or a strain of the plant, the basic unit in forestry is generally the indi-
vidual, and an infinite number of genotypes is involved in each population (Painter 1966).
As long as the genetic base in forest plantations is not allowed to become too narrow
through vigorous selection in limited populations and through inbreeding, problems are
unlikely to develop on the same scale as in agriculture. Broadly based genetic reserves
may be an additional safeguard for the future; by this means the capacity of the host
populations to react to the changes in pathogen populations may be preserved (Dinoor
1975).
PRACTICAL IMPLICATIONS
Although breeding for disease resistance in forest trees is a long-term project,
success stories are numerous.
Diseases attacking the leaves, limbs and stems of trees have been more thoroughly
investigated than e.g. root rots. Although root rots are economically important, evalua-
tion of resistance, screening of selections and progenies, and the development of effec-
tive inoculation methods cause difficulties when dealing with internal damage. In addi-
tion, root rot fungi usually invade the dead heartwood of the tree, in which the possi-
bilities of introducing changes leading to resistance seem smaller than in living
tissue. There have, however, been some positive results from altering by selection the
toxic substances present in the heartwood (Bjorkman 1966; Cech ejt l. , 1966).
Two examples, one entailing a change of species and subsequent selection, the
other selection and further breeding, may serve to illustrate how the techniques on
breeding for disease resistance outlined above can be successfully applied in practice.
Cupressus macrocarpa Hart, and . Lusitanica Mill were introduced into East Africa
in the early J900s. C. macrocarpa wag favoured in plantations because of its somewhat
faster growth. However, problems were soon encountered with cypress canker, caused by
the fungus Monochaetia unicornis (Cook & Ellis) Sacc. The pathogen was studied inten-
sively in the early 1950s. It was found that it occurred on the indigenous species
Juniperus procera Hook, in which it caused little damage. Three strains of the fungus
were isolated, two unimportant and one virulent one; these were brought into culture to
facilitate their study. Then an effective inoculation technique was developed, involving
wounding the stem before the application of virulent inoculum. It was found that disease
resistance of individual trees could be evaluated by measuring the rate of diameter
increase of bark lesions on the trees over a period of three months after inoculation
Using this method it was demonstrated that C. lusitanica was only half as susceptible
- 184 -
* B C inacrocarpa, and a change of species was therefore recommended and undertaken. The
sane inoculation technique was subsequently used to test open pollinated C. lusitanica
progenies of select plus trees (Dyson 1974). Some 80Z of the total reafforestation
area in Kenya is presently planted with C. lusitanica. All seed used originates in seed
stands selected for good form, high yield and a high degree of disease resistance.
Western white pine (pinus nonticola Dougl.) is among the most valuable timber trees
in USA. Blister rust, Cronartium ribicola Fisch. y was first accidentally introduced into
USA in the 1920s, and by 1941 the disease had reached epidemic proportions. Observations
in natural populations of the pine showed that variation existed in disease resistance.
Apparently resistant or immune p he no types were selected, and resistance was tested by the
response of grafts and progeny of selected trees over several years and at several sites.
As the alternate host for the disease is Ribes spp. this plant was interplanted in the
experimental plots to augment inoculum potential. Results from grafts and progeny trials
indicated that both additive and non-additive gene effects were involved in resistance.
The original selections were screened for high general combining ability by means of these
trials.
The next step in the breeding program involved controlled crosses between promising
phenotypes; these were tested for two years under intense artificial inoculation, using
an "inoculation tent" which maintained humidity at levels favourable to the pathogen.
It was found that no selection yielded progeny that were completely immune to the
disease, and that a wide range of variation was apparent in the level of resistance
transmitted by the selected trees (Bingham et l. I960; Hoff J966; Bingham J969).
Practical breeding programs now utilise natural intra-specific variation in western
white pine. The Forest Service of the Northern Region has embarked on a project to mass-
produce ?2 seed from screened intra-specific Fj hybrids of pheno typically resistant
parents. Heritability studies have indicated a gain of some 20Z of F] stock in disease
resistance under intense artificial inoculation conditions. Practical problems encoun-
tered involve seed orchard technology, size of gene pool needed for each planting area
to assure adequate genetic variation for further improvement of desirable traits, and
procedures to speed up selection and testing of new individuals (Hoff 1966).
During the last ten years there has been a growing interest of inter-specific
hybridization in the white pine group. There are more than 20 species of 5-needled white
pines exhibiting varying degrees of resistance to blister rust. The breeding programs
have concentrated on those 14 species which possess the widest adaption to varying site
and climatic conditions and the strongest species-wide resistance. Comparative world-
wide observations have been made on the relative resistance of the various species, and
an international list of 'tentative rankings 1 has been drawn up. Resistant hybrids
between a number of species have been produced and tested on an experimental scale;
especially some of the Asiatic species (P. griff ithii, ]?. Armandii, . Koraiensis)and
their hybrids have proved promising, exhibiting a much higher degree of resistance than
e.g. P. monticola (Bingham J972).
Steps in developing resistance through inter-specific crosses parallel those pro-
ducing resistant strains within a species (Callaham e_t aJU 4966).
The role of forest genetics will become progressively greater as forestry becomes
more intensive. The growing of big, uniform plantations with stock often bred for non-
fitness traits like growth and stem straightness in the absence of potential pathogens
will increase the disease hazard. Efficient plant breeding programs should include plans
to deal with the threat of undetected pathogens and new virulent races of existing ones
before they can cause excessive damage. As stressed earlier in the paper, this can best
be done by maintaining heterozygosity and a broad genetic base in the plantations as
well as in the genetic reserves established or maintained in connection with the breeding
programs .
International co-operation is essential for rapid progress in disease resistance
breeding. The co-operation should include genetic conservation of natural stands to serve
as gene pools for disease resistance, international gene banks of pest resistant geno-
types for the use in breeding, and an increasing exchange of information.
- 185 -
BIBLIOGRAPHY
Allard, R.W., 1964. 'Principles of Plant Breeding 1 . Wiley & Sons Inc., New York.
Anon., 1974. Commonwealth Scientific and Industrial Research Organization (CSIRO).
Annual Report 1973-74.
Benedict, W.V., 1964. Principles, procedures and problems in controlling forest pests.
FAO/IUFRO Symposium on Internationally Dangerous Forest Diseases and Insects, Oxford
FAO/FORPEST - 64, IX. FAO, Rome.
Bingham, R.T., 1969. Rust resistance in conifers - present status, future needs. Second
World Consultation on Forest Tree Breeding, Washington. FO-FTB-69, 5/2, FAO, Rome.
Bingham, R.T., Squillace, A.E. and Wright, J.W., 1960. Breeding blister rust resistant
western white pine. Silvae Genetica 9: 33-41.
Bingham, R.T., 1972. Taxonomy, crossability, and relative blister rust resistance of
5-needled white pines. 'Biology of Rust Resistance in Forest Trees 1 . U.S. Department
of Agriculture, Forest Service. Miscellaneous Publication 1221. pp. 271-280.
Bjttrkman, E. (1966). Status and trends in research related to the resistance of forest
trees to disease in northern Europe, Gerhold, H.D. e al . (Eds). 'Breeding Pest
Resistant Trees'. Pergamon Press, Oxford, pp. 3-JO.
Borlaug, N.E. (1966). Basic concepts which influence the choice of methods for use in
breeding for disease resistance in cross-pollinated and self-pollinated crop plants.
Gerhold, H.D. t l. (EdsX 'Breeding Pest Resistant Trees 1 . Pergamon Press, Oxford,
pp. 327-344.
Buxton, E.W, (1961). Mechanisms of variation in the pathogenicity of Fusarium oxysporum.
Recent Advances in Botany 1; 502-507.
Callham, R.Z. et_ al . (1966). General guidelines for practical programs toward pest-
resistant trees. Gerhold, H.D. et al. (Eds). 'Breeding Pest Resistant Trees'. Pergamon
Press, Oxford, pp. 489-493.
Cech, F.C. e al^. (1966). Breeding conifers for resistance to Fomes annosus. Gerhold, H.D.
e l . (Eds). 'Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 483.
Cowling, E.G. (1969). Principles of genetic improvement in disease resistance of forest
trees. FAO/N.C. State Univ. Forest Tree Improvement Training Center. School of Forest
Resources, N.C.S.U., Raleigh. Lecture Notes, pp. 196-200.
Day, P.R. (1972). The genetics of rust fungi. 'Biology of Rust Resistance in Forest Trees'.
U.S. Department of Agriculture, Forest Service. Miscellaneous Publication 1221. pp. 3-17
Dinoor, A. (1975). Evaluation of sources of disease resistance. Frankel, O.H., and
Hawkes, J.G. (Eds). 'Crop Genetic Resources for Today and Tomorrow'. IBP 2. Cambridge
University Press, pp. 201-210.
Dyson, W.G. (1974). Breeding for disease resistance. Report on the FAO/DANIDA Training
Course on Forest Tree Improvement, Kenya. FAO/DEN/TF 112. FAO, Rome. pp. 292-298.
Franz, J.M. (1964). Forest insect control by biological measures. FAO/IUFRO Symposium on
Internationally Dangerous Forest Diseases and Insects, Oxford. FAO/FORPEST-64, IX,, FAO
Rome.
Gibson, I.A.S. (1975). 'Diseases of Forest Trees Widely planted as Exotics in the Tropics
and Southern Hemisphere 1 . Commonwealth Mycology Institute /Commonwealth Forestry
Institute, Oxford.
- 186 -
Gilaour, J.W., and Vanner, A.L. (1971). Radiata pine needle blight (Dothistroma pini) .
Fungicide and Nematocide Tests 27. Am. Phytopath. Soc.
GSumann, E. (1950) . 'Principles of Plant Infection*. Crosby Lockwood & Son, London.
Heimburger, C. (1962). Breeding for disease resistance in forest trees. For. Chron.
33:356-362.
Heimburger, C. (1966). Susceptibility to a serious fungus attack as a genetic barrier
between aspen species. Gerhold, H.D. e jalU (Eds.) 'Breeding Pest Resistant Trees'
Pergamon Press, Oxford, pp. 391-394.
Heybroek, H.M. (1969). Three aspects of breeding trees for disease resistance. Second
World Consultation on Forest Tree Breeding, Washington. FO-FTB-69, 5/4. FAO, Rome.
Hoff, R.J. (J969). Blister rust resistance in western white pine. Gerhold, H.D. et al.
(Eds). 'Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 119-124.
Lagner, W. (1952). Reziprok unterschiedliches Verhalten von Larchenbastarden gegen eine
Nadelerkrankung. *L. Forst genet ik ^: 78-81
Luig, N.H. (1971). 'Strong' and 'weak' genes for stem rust resistance in wheat breeding.
Aust. PI. Breeding Conf . Perth, W.A. 8-3.
Mclntosh, R.A. (1971). Wheat genes for rust resistance in space and time. Aust. PI.
Breeding Conf. Perth, W.A. 8-5.
McNabb, H.S. Jr (1964). A 'new* concept of forest tree disease control: physiological
suppression. FAO/IUFRO Symposium on Internationally Dangerous Forest Diseases and
Insects, Oxford. FAO/FORPEST-64, IX. FAO, Rome.
Neely, D. (1975). Sanitation and dutch elm disease. Burdekin, D.A., and Heybroek, H.M.
(Eds). 'Dutch Elm Disease'. Proceedings of IUFRO Conference, Minneapolis 1973. USDA
FS Northeastern Forest Experiment Station, Upper Darby, PA. pp. 76-87.
Painter, R.H. (1966). Lessons to be learned from past experience in breeding plants for
insect resistance. Gerhold, H.D. t al (Eds). 'Breeding Pest Resistant Trees'.
Pergamon Press, Oxford, pp. 349-355.
Palmberg, C. (1969). Femes anno s us Fr. (Cke.) - a universal problem. Silva Fennica
3:33*49.
Qualset, C.O. (1975). Sampling germplasm in a center of diversity: an example of disease
resistance in Ethiopian barley. Frankel, O.H., and Hawkes, J.G. (Eds). 'Crop Genetic
Resources for Today and Tomorrow'. IBP 2. Cambridge University Press, pp. 81-98.
Rudd-Jones, D. (1954). Studies on a canker disease of cypress in East Africa. Ann. App.
Biol. 4J: 325-335.
Schreiner, E.J. (J&66). Future needs for maximum progress in genetic improvement of
disease resistance in forest trees. Gerhold, H.D. e jal_. (Eds). 'Breeding Pest Resistant
Trees'. Pergamon Press, Oxford, pp. 455-466.
Scegaard, B. (1966). Variation and inheritance of resistance to attack by Didymascella
thu j ina in western red cedar and related species. Gerhold, H.D. et al. (Eds).
' Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 33-38.
Tarr, S.A.J. (1972). 'The Principles of Plant Pathology' Macmillan Press, London.
Van der Plank (J963). 'Plant Diseases: Epidemics and Control'. Academic Press, New York,
London .
Watson, I.A. (1971). Breeding for disease resistance. Aust. PI. Breeding Conf. Perth,
W.A. 8-37.
Wood, F.A. (1966). The current status of basic knowledge of forest tree disease resistance
research. Gerhold, H.D. l (Eds). 'Breeding Pest Resistant Trees'. Pergamon Press,
Oxford, pp. 293-300.
- 187 -
ECONOMIC CONSIDERATIONS IN FOREST TREE BREEDING PROGRAMMES
B. Ditlevsen
National Forestry Service, Denmark
TABLE OF CONTENTS
Page
Introduction 187
Market factors 187
Choice of procedure 188
Evaluation criteria for a genetic improvement plan 189
Costs and benefits 189
Costs 190
Fixed costs 190
Variable costs 191
Benefits 191
Dubious points in evaluation 192
Optimization of improvement programmes 193
Final comments 194
Bibliography 195
INTRODUCTION
Tree breeding through genetic manipulation is one of several methods available for
increasing forest production and the efficiency of processing procedures. The potential
benefits and costs of tree breeding must therefore be considered within the overall frame-
work of the task of reforestation and utilization as a whole.
Restocking programmes may be promoted by individuals, private companies, ministries
or other public bodies. If the capital comes from private sources, the usual objective
is to maximize annual benefits. Projects promoted by public bodies may be established for
social purposes, but also to obtain direct economic benefits. In the latter case, the
objective will be to achieve maximum production rather than to maximize the economic
benefits, in order to ensure that the country has an adequate supply of wood, or to deve-
lop (or maintain) a strong forest products industry.
In this paper we shall deal only with considerations relating to financial cost-
benefit models, but the breeder must bear in mind that an analysis of social costs and
benefits may be necessary in certain cases, and may lead to entirely different conclusions
(Reilly, 1977).
MARKET FACTORS
The market for genetically improved seed is like that of other commodities. Growers
supply seed in accordance with existing demand.
The demand for genetically improved seed depends on the nature and level of the
demand for the commodities grown from the seed, and this, in turn, reflects the demand for
- 188 -
products made from these. This is of great importance, since the price that the buyers
are willing to pay for the improved seed will be equal to the price of the products grown
and manufactured from this seed, less the costs of cultivation and processing.
Since the benefits of a tree breeding programme will extend beyond the first few
years, it is advisable to have some idea of future trends in the market for the processed
products. The level of demand forecast will influence the amount of seed to be produced,
while the nature of the demand will influence the characters to be selected and th intensity
of selection.
In a stable market situation in which supply and demand, and hence prices, do not
vary, the point at which the supply of genetically improved seed is in balance with demand
will be mirrored in the market price of the seed. If the amount available on the market
is excessive and demand does not change, the price will fall, but will eventually return
to its original level as supplies are reduced. An increase in the demand for improved seed
owing to an increase in the annual planting rate will result in a rise in the price.
This very simplified description of the market relates to the prices of a surplus
offered to other seed growers. If the surplus seed production of the organization's
programme represents a considerable proportion of the total market, the price will fall.
This must be borne in mind when evaluating the seed. Likewise, if the organization decides
to buy genetically improved seed from other growers instead of starting its own programme,
the increased demand may be sufficient to provoke a considerable rise in prices. However,
if the effect on supply and demand is not very great, no great changes in price will occur.
Markets for genetically improved seed are sometimes highly specialized* Breeding
programmes cover a wide range of species, selected characters and selection intensities.
The market is often dominated by one or a few producers, and the price will be based, not
on market forces, but on the cost of production plus a margin of profit. The cost of
producing genetically improved seed on the basis of one's own programme may be very different
from the market price, or the price that an organization was willing to pay for the seed.
CHOICE OF PROCEDURE
There are two kinds of alternative procedures or strategies: those arising from the
decision on whether or not to undertake the programme, and those arising from having
chosen the improvement programme. The strategies in the first category are as follows:
1st to continue using unimproved seed;
2nd to buy genetically improved seed from other sources;
3rd to participate in a cooperative tree breeding programme;
4th to undertake one's own genetic improvement programme.
Before undertaking a comparison of these strategies, it is necessary to ascertain
whether it is technically possible for the organization to undertake a genetic improvement
programme. This will depend essentially on the existing demand for genetically improved
seed, the resources available to the organization, the size of the programme proposed, and
the objectives of the directives. The resources available include not only highly skilled
breeders and other labour, but also capital, land and, above all, the size and nature of
the population available for selection purposes. Research may also prove necessary to
establish flowering models, vegetative propagation techniques and other operational proce-
dures for seed orchards.
The kind of genetic improvement programme will depend mainly on the available data
on the heritability of important improvement characters, the additional genetic gain at
different levels of selection intensity, the genetic relationship between various characters,
etc. The size of the programme will depend on the organization's seed requirements and
the demand for seed surpluses*
Initially, most organizations will not have sufficiently detailed information to
undertake big programmes, and analysis will be limited to a comparison of these progr
- 189 -
with one or another of the above alternatives* As the programme proceeds and more is known
about the genetic improvement characters of the species, more advanced strategies may be
contemplated.
EVALUATION CRITERIA FOR A GENETIC IMPROVEMENT PLAN
Economic experts have not yet completely solved the problem of choosing the right
economic criterion for comparing the benefits and costs of alternative procedures over a
period of years, or for accepting an individual project. But most now agree that it is
necessary to have some method of discounting to reduce benefits and costs of different
dates to a common denominator at any given moment. A dollar is worth more to an organi-
zation today than at a future date, because it can be used to produce additional income.
The following two evaluation criteria are the ones most frequently used:
1 The current net value of a project is defined as the difference between the costs
and benefits attributed to the project actualized at the appropriate rate. If the actual
net value is positive, the project can be considered economically feasible and may be
accepted. If there is more than one project, preference should be given to the one which
offers the highest positive actual net value.
2. The rate of internal yield is the type of actualization which gives an actual net
value equal to zero, i.e. the actualized value of the benefits is equal to the actualized
value of the costs. An individual project may be accepted if the rate of internal yield
proves higher than the minimum acceptable rate adopted for investment purposes. The choice
between projects wil 1 be determined in favour of the project with the highest rate of
internal yield.
In some cases, the project that offers the highest actual net value does, not
necessarily offer the highest rate of internal yield. In such a case, the criterion of
the actual net value is the one that should be followed.
The actual net value of a programme (CNV) may be expressed in algebraic form through
the following equation:
CNV = f2 (B i- - O/C 1 + O't where:
t=0 C L
B denotes the benefits generated by the programme in the year t,
C t denotes the coasts incurred by the programme in the year t,
i denotes the rate of interest for actualizing benefits and costs, expressed as a decimal.
In calculating the rate of internal yield it is not necessary to have any predeter-
mined rate for actualization purposes. In order to decide whether the programme should be
accepted or not, one needs an estimate of the minimum rate of yield acceptable for the
organization contemplating the investment. This rate will be equal to that which the
organization should have earned in alternative investments, i.e. the in timely placement
cost of its capital.
Most privately-owned organizations expect new investments to generate as a minimum
a rate of profit equal to the weighted average profits from other sources (capital shares
and loans, undistributed profits and external investments), according to the relative
proportions of these resources used by the organization.
EVALUATION OF COSTS AND BENEFITS
The simplest method of analysing costs and benefits is to limit the analysis to
actual cash transactions. The costs will therefore include wages and salaries, the price
of purchasing land and other fixed aspets, such as buildings, plant and equipment, and
the costs of the raw material. Values must be assigned to the fixed assets and land at
the end of the time horizon and included as benefits.
- 190 -
Two aspects of the cost /benefit ratio must be examined. At what point in the pro-
duction process should the costs and benefits connected with the genetic improvement pro-
gramme be evaluated? And how should they be planned over the years?
The costs must be charged to the account at the time they occur. The benefits are
usually evaluated at the time the seed is collected. All the costs relating to the selec-
tion of parent trees and the establishment and management of the seed orchard are naturally
included in the analysis, while those relating to progeny and provenance trials, hybridiza-
tion programmes, etc., are neither partially nor fully attributed to the main improvement
programme, and it may therefore be necessary to earmark funds for these items in some way.
It is preferable that this be from profits generated by the different programmes. Some-
times, however, it is necessary to adopt an arbitrary procedure based on the breeder's
own assessment.
The value of genetically improved seed is based on its price on the free market
(if there is one) or the value of profits generated by sale of the products obtained from
the seed using a method called residual value price. This implies 'deduction of the costs
of cultivation, harvesting, processing and distribution from the price of an end (or
intermediate) product derived from the seed. The prices of sawnwood and pulp are appro-
priate starting points for an evaluation of the benefits of a forest tree breeding programme,
COSTS
Fixed costs
As in any other form of technological change, the costs of tree breeding may be
fixed, or may vary in accordance with the size of the seed harvest produced. They are
fixed if they are not affected by the size of the harvest, and variable if they vary in
direct accordance with it.
Selection and progeny trials of parent tree for the seed orchard constitute the
main sources of fixed costs connected with the management of seed orchards. This is
because the number of phenotypes selected to be propagated in the orchard is normally
independent of the size of the orchard.
1 Cost of selection per parent tree
The cost of selection per parent tree for a seed orchard depends on the selection
intensity and the criteria or characters on which selection is based. Van Buijtenen and
Saitta (1972) showed that the initial selection cost per tree tends to increase at a
relatively more rapid rate as selection intensity increases. Per example, the cost of
initial selection per tree at high selection differentials of 3 or 4 standard deviations
above the average was almost five times greater than with relatively little deviation
from the standard.
The cost of initial selection also increases greatly in proportion to the number
of characters to be selected. For a selection intensity of one percent and assuming that
the characters are not correlated and that the costs of the search amount to $1 per 100
trees studied, the selection cost would be as follows:
Number of characters Number of trees studied Cost per parent tree
to be selected per parent tree selected ($)
1 100 1
2 10 000 100
3 1 000 000 10 000
The selection cost increases 100 times for each additional character selected.
- 191 -
2 Progeny trials
A progeny trial may represent one of the most expensive aspects of the genetic im-
provement programme, and its costs are very difficult to figure. If progeny trials are
designed to provide information for the purpose of roguing a first-generation seed orchard,
the total cost of the trials will be assigned to this phase of the improvement programme.
If they are also designed to provide the basis for a second-generation selection, a method
is required for assigning the costs between the first and second generation orchards.
The ideal would be to base cost assignment on the profits generated by each generation,
but it may prove impossible to quantify these. In this case, a general factor is needed
to cover most circumstances, such as the relative costs of selection.
3 Other fixed costs
The annual costs of administration and research constitute the other most important
categories of fixed costs* The costs of administration may be assigned pro rata among the
costs of wages, salaries and materials directly generated in each phase of an improvement
programme, but it is more convenient to treat them as a fixed annual cost.
If the research programme is not designed to benefit the improvement programme
carried out in combination with it, the costs must be changed to the future improvement
programmes on behalf of which they are designed, when these programmes are produced.
Variable costs
The remaining costs of establishing and managing a seed orchard are closely connected
with its area or size. For any management system, they also depend directly on the seed
yield. They include the costs of preparing the site, planting, vegetative propagation,
fertilization, pruning, roguing, overall maintenance and protection.
Establishment costs will vary according to the number of clones, the method of
propagation and the initial density of the forest. The per hectare seed production of a
species depends on the extent of site preparation, particularly the amount of fertilizer
used, and the density of the forest. Keiding (1975) has given a detailed summary of the
costs of establishing seed orchards.
The larger the area of the seed orchard, the lower the costs per kilogramme of seed
produced. However, a moment may come when the orchard is too large and diseconomies of
size or scale become apparent, owing mainly to a disproportionate increase in the costs
of administration, or perhaps because the collection of graft material is very expensive
or more parent trees are required.
Thus the production cost of seed in an orchard will follow the well-known U-shaped
curve of production theory, i.e. it will fall at the beginning as the area of the orchard
increases, and rise at the end as the fixed costs start to rise. Similar relations may
be observed in individual operations, particularly between the average cost of fertilizers
per kilogramme of seed produced and seed production per hectare; and between the spacing
of ortets and seed production per hectare.
BENEFITS
Tree breeding can benefit forest production in four ways: first, by increasing the
yield of wood per hectare ("yield effect 11 ); second, by inducing a rise in the prices of
the different products harvested ("price effect 11 ); thirdly, by reducing costs ("cost re-
duction effect"); and fourthly, by decreasing the economic rotation for the forest
("rotation effect").
1 Yield effect
This effect can be obtained by cultivating genetically improved trees which give
a higher yield, or by reducing losses in yield by cultivating trees better adapted to the
environment, e.g. trees more resistant to wind in areas with cyclones or trees resistant
to diseases, insects, damage produced by frost or snow or drought* One of the problems
in trying to increase the inherent speed of growth, however, is the relatively low heri-
tabllity of this character.
- 192 -
2 Price effect
If we assume that the price of the improved seed will include any economic surplus
from harvesting and processing of the trees grown from the seed, the following factors
will influence the value or price of the seed:
a) Higher yield of processed articles. For sawn logs, longer trees, with straighter
trunks and improved form, and less compression wood and twisted grain, provide
a higher yield of saw timber per unit of roundwood volume. For pulpwood, a
higher yield is obtained from trees with greater basic density, little compression
wood and better fibre properties. A reduction in the variability of these pro-
perties of the wood also produces a higher yield.
b) Improvement in the class or quality of the processed articles. Higher classes
of saw timber, with consequently greater value, can be achieved by reducing the
size of the branches and the amount of twisted grain and compression wood,
improving branch angle, etc. Higher classification of the pulp can be obtained
by cultivating trees with appropriate fibre properties.
c) Reduction of felling costs. These costs can be reduced through higher yields
per hectare, larger average trunk size at time of felling, wider spacing at
time of planting, and better branch characteristics (reduction in snedding time).
d) Reduction of processing costs. For saw logs, straighter, larger trees of im-
proved form and uniform size reduce sawing time; improved wood properties can
reduce drying time in the sawmills and cooking and grinding time in the pulp
plants.
3 Cost reduction effects
Faster-growing, better quality trees enable forestry directors to space tree plantings
more widely, which means fewer seedlings per hectare and easier access. Many operations
are thus shortened.
* Rotation effect
Higher prices and yields and lower cultivation costs mean economic rotation can be
considerably reduced. This also offers the advantage of a reduction in the amount of land
necessary.
The benefit side of the budget is the most difficult aspect in any genetic improve-
ment programme. How can the benefits deriving from the programme be evaluated? As we
have already indicated, evaluation can be done in two ways: one based on the market price
of the seed produced, and the other based on the residual value price method.
If the market price method is adopted - which is the exception rather than the rule,
since the improved seed, even of the same species, is rarely of the same genetic quality -
the problem is to identify any change in price deriving from the increase in the supply
of seed.
If the residual value method is adopted, the procedure used will depend on the
products cultivated or processed from the seed of the seed orchard and the characters
selected.
The benefits to a plantation from genetic improvement can be evaluated in terms of
the seed produced, actualized to the time the seed was collected and from the time profits
begin to accumulate, using data relating to the initial planting density, the gerrai native
capacity of the seed and the rate of roguing in the nursery to express the benefits in
terms of dollars per kilogramme of seed.
DUBIOUS POINTS IN EVALUATION
Most studies imply that physical production inputs and outputs and their prices can
be estimated without error. But this is not so.
- 193 -
An analysis which takes into account the doubts inherent in probability distribu-
tions for basic inputs and outputs and their values is much more exact with regard to the
collection and analysis of data than one based on average or probable values. Also, it
is very often impossible to make reasonable subjective estimates of the probability distri-
bution of many variables, particularly in the analysis of investments that imply techno-
logical changes, such as the genetic improvement of trees, in which even the average increase
in benefits is very often unknown.
One possible way of resolving this difficulty is to assume that each option evaluated
will have the worst possible result, and to select the one that gives the highest value
for the economic criteria selected. This is a highly conservative method and may result
in projects with high potential benefits not being considered.
In projects with investments of a biological type the costs can usually be evaluated
fairly accurately, but the benefits raise difficulties. Tree breeding is a project of
this kind. Lundgren and* King (1965) and Davis (1967) resolved the problem by calculating
the levels of genetic gain that should be obtained in a tree-breeding programme based on
known costs and then comparing these gains with data available from progeny trials. Another
method is to limit the studies to gains that can be evaluated easily, as, for example, in
the analysis by Swoford and Smith (1971) of the national forests in southern U.S.A.
Finally, it may prove more satisfactory to make a sensitivity analysis of the key
parameters, such as prices, or of those which are the object of subjective evaluation or
are subject to big fluctuations. This requires an evaluation of each project on the basis
of the most probable floor and ceiling values of the key variables. The main objective
of the analysis is to judge whether it is to be assumed that a project is unacceptable
in terms of its assumed floor, or whether the order of the competing projects should be
modified.
OPTIMIZATION OF IMPROVEMENT PROGRAMMES
According to the traditional production theory, the optimum result of a project is
defined by the point at which the marginal cost is equal to the price of the production,
where the marginal cost is defined as the cost of producing an additional production unit.
For a project with a life of one year or less, this is relatively easy; but for long
projects like tree-breeding programmes, the situation is complicated by the additional
variable represented by time.
The simplest solution to this problem is to assume that there are no restrictions
on resources and then submit all the feasible genetic improvement options to separate
cost-benefit analysis. The one that generates the maximum economic benefits will be chosen.
Options may vary according to the number and nature of the characters to be selected,
selection intensity, the number of plus trees propagated in the seed orchard, its area and
the management system adopted. For each set of characters to be selected, there will be
a range of production possibilities based on the above factors.
This method may prove extremely complicated and require much time as the number of
feasible improvement programmes increases. Owing to inadequate data this very rarely
happens; but if sufficient information is available, mathematical programming techniques
may prove appropriate.
Linear programming was used by van Buijtenen and Saitta (1972) to derive optimum
solutions based on the size of the harvest, the area of the seed orchard and whether or
not the orchard had to be rogued. The technique is very appropriate for the optimum ear-
marking of scarce resources between competing requirements, in terms both of individual
programmes and of time. The process can be performed rapidly and exactly by computers.
A more recent development is Porterf ield's study (1974) on potential gains in genetic
improvement programmes for Pinus taeda in the southern U.S.A. using goal programming. This
is simply a modification of lineal programming, but here the restrictions are replaced by
the goals. Many goals can be included in this technique, while there is only one in lineal
programming, i.e. maximization of the net returns in a single objective function. Characters
which it was considered would affect the volume and specific weight were incorporated in
the model. Only characters affecting the volume were taken into consideration for sawlogs,
but both characters were taken into account for pulpwood.
- 194 -
The main advantage of the model is that the goals can be modified, absolutely or
relatively, according to changes in the market. It can provide for the genetic selection
reaction of indigenous populations, roguing and progeny trials. By specifying the selection
intensity for a given character and the percentage of improvement desired, a solution is
obtained which gives the minimum deviation from the given genetic goal subject to restric-
tion under the capital expense budget.
9 The purposes of any optimization will depend on the organization contemplating the
investment. In view of the rapid changes in the market for wood products, many organiza-
tions prefer to adopt a conservative genetic improvement strategy which maintains maximum
flexibility. Greater importance may be attributed to characters that may be important for
various generations in a range of different environmental and economic conditions. As a
general rule, an increase in the yield of wood per hectare and greater resistance to diseases
are two characters of this kind, while it may be assumed that wood properties such as grain
are not so important.
Other organizations are more concerned with improving first-generation characters
which are obviously inferior but which might provide large gains rapidly, and concentrating
more on other characters in later generations.
Great care must be taken in applying very advanced optimization procedures, since
they depend entirely on the data on which they are based. The benefits or gains deriving
from the tree breeding constitute a major problem. Very little is yet known of the gains
obtained through genetic improvement of trees in a complete rotation of a plantation re-
generated by genetically improved seed, particularly in cases in which other cultural
techniques, such as pruning and thinning, are used. For plantations that produce only
one product, e.g. pulpwood, have a short rotation and are not thinned, the gMns can be
forecast with more accuracy than for those which are regularly thinned and which produce
both logs and pulpwood in long rotations.
FINAL COMMENTS
Genetic improvement is an expensive programme that must be justified by its potential
benefits.
Nikles (1973) gives a series of examples of the economic considerations relating
to the genetic improvement of forest trees.
Breeders usually choose the improvement programme which provides the biggest and
most rapid gains and has the shortest improvement cycle.
Smith and Zobel (1974) have shown that there will be enormous losses in potential
gains if the source of the seed is unknown, and it is important that this source be con-
sidered and thoroughly investigated before a genetic improvement programme is formalized.
They also considered that important gains in straight ness of trunks, specific
weight, length of tracheid and resistance to rust can be confidently assumed, as they are
highly inheritable.
As regards costs, they proved that, when all the costs of a representative seed
orchard were updated to a common point in time, the selection of parent trees with the
subsequent progeny trials represented only 10 percent of the total costs. Hence an
improvement in selection standards may produce only a relatively small increase in costs.
This bears out the results of van Buijtenen and Saitta (1972), who showed that selection
was the most effective phase in the tree-breeding programme they were studying, in terms
of the profitability of investment in a first-generation seed orchard. Most (89 percent)
of the total costs of the seed orchard programme concern preparation of the site, cost of
the land, supervision, fertilizers, protection, harvesting and extraction of seed, and it
is important to realize that these costs do not vary with the genetic quality of the tree
selected.
Some characters may compete with each other. This seems to be characteristic of
characters of quality, such as straightness of trunk and branches, when these are selected
in conjunction with fast growth.
- 195 -
The profitability of tree breeding is considerably increased by progeny trials and
roguing of the seed orchard (Smith and Zobel, 1974). Consequently, if it is desired to
obtain maximum economic benefits, progeny trials and high selection standards must be used.
The seed yield per hectare of the seed orchard must also be maximized through appro-
priate use of fertilizers and proper spacing.
Owing to the variety of species and the doubts regarding benefits, it is difficult
to generalize about the most appropriate method of formulating a tree improvement programme.
Perhaps all we can do is to repeat the conclusion of Smith and Zobel (1974) that the most
profitable tree improvement programmes usually have the following characteristics:
1 One species is grown over a wide area;
2 The desired characters have moderate to high heritability;
3 Crossing of the species can be easily controlled, and the species is a prolific
seed producer;
4 The values of the wood products are high and it is to be assumed will remain
high.
BIBLIOGRAPHY
Davis, L.S., 1967. Investments in loblolly pine clonal seed orchards: production costs
and economic potential. J^ For* 65.
Keiding, H. , 1975. Economic Considerations in IMproved Seed Sources, FAO/DANIDA Training
Course on Forest Seed Collection and Handling, Chiang Mai, Thailand, 1975.
Lundgren A.L., and King, J.P., 1965. Estimating financial returns from forest tree
improvement programs. Soc. Amer. For. Proc.
Nikles, D.G. , 1973. Economic aspects of tree improvement. Cost benefit of tree-breeding
programmes. FAO/DANIDA Training Course on Forest Tree Improvement, Kenya.
Porterfield, R.L., 1974. Predicted and potential gains from tree improvement programs -
a goal programming analysis of program efficiency. N.C. State Univ. Tech. Rep. 52,
Raleigh, N.C.
Reilly, J.J., 1977. Economic Aspects of a Tree Improvement Program, International Training
Course in Forest Tree Breeding, Canberra, Australia, 1977.
Smith H.D., & Zobel, B.J., 1974. Genetic gains and economic considerations. Tree Improve-
ment Short Course, School of Forest Resources, N.C. State Univ., Raleigh, N.C.
Swofford, T.F. & Smith, O.D., 1971. An economic evaluation of tree improvement in the
southern national forests. U.S.D.A. For. Serv. South Reg. Publ. 10.
Van Buijtenen, J.P. & Saitta, W.W. , 1972. Linear programming applied to the economic
analysis of forest tree improvement, J. For. 70/3.
- 196 -
Participants visiting the CON ARE Pinus caribaea plantations in Chaguaramas.
- 197 -
PLANNING AND STRATEGIES OF A TREE IMPROVEMENT PROGRAMME 1
Palmberg, Paul and Willan
CONTENTS
Page
Introduction ................................................... 198
Preliminary considerations ......................................... 198
Administrative matters .......................................... 198
Technical questions ............................................. 198
Selection criteria and the basis for selection .......................... 199
Selection criteria ............................................. 199
Basis for selection ............................................ 199
Strategy ...................................................... 200
General principles ............................................. 200
Populations .................................................. 200
Considerations in choice of breeding methods ......................... 201
Breeding methods and their application .................... ......... 202
A dynamic breeding programme ....................................... 202
Current seed requirements ....................................... 202
Base for future selection ....................................... 203
Progressive improvement of seed through successive orchards and clone banks ... 203
Conservation of gene resources .................................... 203
Demonstrating progress .......................................... 203
Staff organization, facilities and administration ........................ 204
Staff and facilities .......................................... 204
Revisions of plans and publication of results ... . . ..... . ............. 204
Summary ...................................................... 204
Bibliography .................................................. 205
Annexes 1-6 Strategy model for Improvement of specific species
This lecture is a revision of the one originally presented by Or D.G. Nikles at the
FAO/DANIDA Training Course in Kenya in 1973, and published in the Report of that
Course (FAO/DEN/TF-112. FAO t Rome 1974).
- 198 -
INTRODUCTION
Planning a tree improvement programme involves many considerations. A primary factor
is the stage in development of the reforestation project. A common situation, and the one
adopted for reference in this lecture, is where a species with considerable known or
suspected geographic variation is already being successfully planted on a moderate scale.
The aim of the reforestation project is to meet the needs for a variety of products for
local use and/or export. The scale of operations is sufficient to justify a continuing,
multi-generation improvement prog ran
Many more and less advanced variations of this reference model can be readily
envisaged, and details of appropriate associated tree-breeding programmes would vary
accordingly. It is not possible to consider many such situations here, so only the
principles of planning a tree improvement programme are emphasized in this lecture.
PRELIMINARY CONSIDERATIONS
The first step in planning a tree improvement programme is to ascertain the types
of products likely to be required and the aims of forest management now and in the future.
Such an investigation and consideration of the following points will be needed in
order to develop a strategic plan for the tree improvement programme: at step two the
following are examined.
A. Administrative considerations
(i) Formulation and statement of the aims of the tree improvement programme in
order that they will contribute most effectively to the overall objectives. Aims must be
expressed as simply and precisely as possible.
(ii) Assurance of possibilities to provide funds, equipment, facilities, and
qualified personnel. Provision for extra training may be required.
(iii) Assurance of programme continuity. It is desirable that suitable personnel
be encouraged to make a career of tree breeding. If there are doubts about continuity of
competent staff, it is essential to choose simple, robust strategies for the programme.
(iv) Possible reorganization of administrative resources to locate tree breeder
and staff in suitable headquarters together with colleagues working on silviculture, soils
and nutrition, wood quality and products, forest management, etc.
(v) Consideration of cooperation in improvement work at local, regional, national
and international levels. This can lead to cost sharing of technical assistance and
research, exchange of ideas and generation of enthusiasm.
B. Technical considerations
(i) Determination of the factors limiting forest production in the region and
ways and means of manipulating these to meet stated overall objectives. It is essential to
build a tree improvement programme on a base of sound silviculture, management and utili-
zation. The relatively long time scale of tree improvement work and the possibility of
rapid technological change in other techniques of forest management and utilization should
also be borne in mind.
(ii) Choice of species and provenances to provide the desired type of products.
Identification of the best species and provenances for each major site type in potential
planting areas is essential.
The availability of basic biological information on the species (i.e. their ecologi-
cal and morphological variability, individual variation and ease of regeneration by seed
and vegetative means), as well as of possible techniques for improvement based on
experiences in other countries should be investigated. Special characteristics of indivi-
dual species can often be used to advantage.
- 199 -
(ill) Determination of the characteristics which are best manipulated by genetic
means* Basic studies to secure reliable estimates of genetic parameters will provide
information of great value in improving efficiency of selection and of breeding strategy.
The development of simple assessment methods, efficient record keeping and data
handling procedures will often be required in connection with these studies.
(iv) Specification of the number of generations and number of years needed to
achieve a certain degree of improvement in important characteristics* This information should
be critically examined in the light of urgency for the improved material.
(v) The gathering, throughout the programme, of data for the evaluation of costs
and benefits.
Evaluation of costs and returns have been discussed in the literature i.e. by Van
Buijtenen (1975); Porterf ield (1978) : Reilly and Nikles (1978); Reich and Carlisle (1978);
and van der Meiden (1978).
SELECTION CRITERIA AND THE BASIS FOR SELECTION
Selection Criteria
In choosing characteristics to be included in the programme, one should concentrate
on a few traits which have high potential for economic gain and which cannot be more
cheaply improved by cultural or technological means. Consideration should be given to
inclusion of a trait if one or more of the following factors are relevant.
(i) economic value is high;
(ii) the trait is likely to be of continuing value even if product demands change
in the future.
(iii) variability and heritability are high, giving potential for considerable
genetic gain;
(iv) the trait is positively correlated with, or is independent of, other
desirable traits;
(v) genetic and management and non-genetic improvement methods interact favourably,
or combine positively to improve the trait.
Some traits are desirable for virtually all end products, and they may be expected
to retain their prime importance fora long time. These are good health, high growth rate,
lack of malformation, straight stem, large branch angle. Several other traits are of great
value for specific products, e.g. special wood properties, branch pattern, branch diameter,
branch longevity, and stem defects such as cone holes.
In planning a programme, attention must be given to ranking the possible selection
criteria in order of likely economic importance many years hence, and adjusting this
ranking for expected genetic improvement. Each tree selected as a candidate for the breeding
programme can then be assessed for overall merit.
The Basis for Selection
The degree of genetic improvement in the desired traits through selection and breeding
is dependent on three components. These are (i) variation; (ii) heritability and
(iii) proportion of trees selected. The degree of improvement is expressed by the Genetic
Gain, AG, which is computed as the heritability x selection differential.
'(i) Variation in the traits to be improved is a prerequisite and a starting point
for any programme. Initial selection should be done in large populations in which the genetic
base has not been narrowed down by e.g. collecting the seed for the establishment of the
stand from only a few trees.
- 200 -
(ii) Heritabtlity (h ) is a measure of the degree to which a character
influenced by heredity as compared to environment, and is expressed as the proportion of
genetic to genetic plus environmentally caused (phenotypic) variation (see paper on
quantitative genetics for more detail)* If selection is carried out in even-aged, regularly-
spaced stands growing on uniform sites, the masking effects of the environment will be
minimised, and accuracy of determining the genetic proportion of the variation will increase.
This accuracy can be further increased by stratifying the plantation into site class areas,
by comparing candidate superior trees to neighbouring trees growing on similar sites,
applying adjustments to total volume for crown size effects, etc.
(iii) Proportion of Trees Selected. The gain component most amenable to influence
by the breeder is the proportion of the population selected. The intensity of selection is
measured by the selection differential (s), i.e. the difference between the mean of the
selected trees for a certain characteristic and the mean of the original population for the
same characteristic. In theory, the larger the selection differential the larger the genetic
gain. In practice, however, to increase the selection differential becomes increasingly
more costly as the area to be searched increases in dimension logarithmically for a given
increase in selection differential (Shelbourne 1973). Thus, a costly search for the 'l-in~a-
million tree 1 will not achieve a correspondingly large selection differential, and less
rigorous selection of individuals for inclusion in the first-round breeding programme may
be called for (say, 0,1-11). In practice, countries with rather small plantations will have
to select even less intensively (probably about 1 percent of the trees).
STRATEGY
Determining the most efficient system of managing the various parts of a breeding
programme within the constraints of time and other resources available can be complicated
and compromises are necessary to resolve conflicts of interest. The scheme that is
developed as a result of all the conditions necessary is termed the breeding strategy. The
process by which the strategy is implemented is termed the breeding method, while the
biological procedures such as grafting, controlled pollination, planting, etc., are the
breeding techniques. Thus, a breeding strategy is an approach to securing the greatest
improvement with the species, resources and conditions prevailing.
General principles
Tree improvement has three major objectives, namely, to provide:
(i) genetically improved seed or other material for immediate use;
(ii) suitable select breeding material for future use;
(iii) adequate genetic information for immediate and future use.
In addition, the need to conserve the gene resources of the original populations and
land races of the species concerned must be recognized.
Populations
The principal objective of the tree breeder is the generation of optimum breeding
populations to create cumulatively better genetypes for advanced-generation seed orchards
(Namkoong 1972). This involves the reduction of the genetic base of the population, thus
creating a serious conflict between short-term and long-term gains through reducing
effective populations site. The conflict can be mitigated by maintaining a hierarchy of
separate populations side by side, representing a series of decreasing selection intensity
but increasing effective population size (Burdon, Wilcox and Shelbourne 1978). These
- 201 -
populations include the following:
(i) Gene pool - a population in which the full range of genetic variation is
maintained; it may sometimes equal the original selection population.
(it) Selection population (base population) - a large population within which
superior trees are selected; about 1 million trees is suggested (500 -
1 000 ha).
(iii) Breeding population - a selected population, containing perhaps 200-300 trees
chosen for their superiority within the selection population. It is used in
whole or in part to generate the next selection population.
(iv) Seed production population - a population of 30-100 trees established to
produce seed for forest plantations (or, for the wood-producing population).
There are many variations of this pattern. Mass vegetative propagation would elimi-
nate the seed-producing population as a separate entity, while a seedling seed orchard
combines (ii) and (iii).
The classic seed-producing population is the clonal orchard, giving very high selec-
tion intensity but still large enough, say 25 unrelated clones, to give genetic "insurance 1 *.
Maximum outcrossing is desired in the seed-producing population although the parents could
be inbred themselves. The breeding population, which would include the seed-production
genotypes, represent a compromise between maximum selection differential and full population
size, but would also allow recruitment of individuals which are superior for particular
traits. A gene pool of standing trees, although existing for gene conservation, cannot be
shielded from all selective processes. Mild and non-specialized silvicultural selection
seems reasonable, while preservation in several contrasting environments should help
maintain genetic diversity (Burdon, Shelbourne and Wilcox 1978).
Considerations in choice of breeding methods
Gains likely to be achieved through tree improvement can be substantial and will be
of a lasting nature. If breeding populations are kept variable and reasonably large (never
fewer than 50 unrelated individuals), steady progress can be made for many generations
through selection.
An understanding of the following biological principles and technical guidelines may
facilitate decisions when carrying out the improvement programme:
(i) Most individuals of most forest species being used in large reforestation
projects suffer depression of growth upon inbreeding. It is therefore essential to maintain
a large genetic base in the selection and breeding populations to allow for mating between
non-relatives.
(ii) With small selection populations, say, only 1 000 individuals, a sacrifice in
gain will be necessary; in order to maintain a large enough population (i.e. selecting a
large proportion of the trees), selection intensity will have to be relatively low.
(iii) Some individuals may have a very high general combining ability (GCA), which
can only be utilized through controlled crosses. Provision for controlled crosses at an
experimental scale should therefore be made at a fairly early stage of the programme.
Choice of appropriate mating designs is of fundamental importance for this step in the
strategy.
(iv) Establishment of early trials is necessary to enable estimation of the effects
of family-by-environment interactions.
(v) Family experiments should normally contain a large number of entries and of
the order of 100 siblings per family. The number of trees per plot should be small and
replications numerous.
(vi) Exploitation of the biological characteristics of the species with ingenuity is
part of the art of tree breeding, for example, by means of clonal propagation, mass
poilination t hybridisation* etc.
- 202 -
(vii) Vhere breeding programmes lack a sufficient number of desirable clones the
situation can be rectified by selecting locally, or importing, additional material for
injection of "new blood" into the breeding population.
Breeding methods and their application
Essentially, there are only two main strategies: one using sexually reproduced
progeny and the other, vegetatively reproduced progeny.
(i) Selection with Regeneration through Seed
(a) Simple or recurrent mass selection. Example: Selection of phenotypically
superior trees, collection of open-pollinated seed and bulking of the seed. Seed from the
original, select trees is harvested continually (simple mass selection), or selection and
seed collection is eventually carried out within their progeny (recurrent mass selection).
The method is cheap, and it can be quite effective where the selection differential is
large and phenotypic selection is accurate (high heritability and environmental effects
minimized within the selection population). It has been used successfully in many countries
and for many species.
(b) Selection and simple progeny testing, with or without further control of
pollination. Example; Culled seedling seed orchards (pollination control achieved by
ing,
ards
isolating the progeny from major sources of contaminating pollen).
(O Selection, full control of pollination, without progeny testing. Examples;
Clonal and control-pollinated seedling seed orchards.
(d) As above but with periodic progeny testing, recurrent roguing and subsequent
selection. Examples; Clonal and control-pollinated seedling seed orchards (a) with
roguing of clones or of families and individuals; or (b) using test results to select
material for new orchards. Note that roguing clonal seed orchards may give limited gain
unless they contain a relatively large number of clones (50+) and the orchards have
originally been established using relatively close planting distances. In some cases, where
specific combining ability has been found to be high, bi-clonal orchards may be established
for seed production.
(e) Hybridization of species or provenances. Examples; Larch species hybrids in
Europe and Japan; P. elliottii and P. caribaea var. caribaea in Queensland. This method is
attractive in some~special circumstances (Brown, 1972).
(ii) Selection with Clonal Propagation;
(a) Without testing. Example; Some poplar and willow culture; . radiata in
New Zealand to a limited extent. ~
(b) With clonal testing. Example; Poplar and willow culture; Cryptomeria in
Japan proposed for P. radiata for future in New Zealand and for Eucalyptus hybrids in Congo.
This is a high gain but "dead end" procedure unless a crossing programme is included to
produce new selection populations periodically.
A DYNAMIC BREEDING PROGRAMME
Current seed requirements. Possible sources of improved seed for immediate use
include the following:
(i) Phenotypic selection and seed collection in local stands of good provenance
generally assures well-adapted populations and gives modest genetic improvement. Genetic
gain is limited by the unselected nature of the pollen parents.
(ii) Collection of seed from thinned seed production areas; ideally, these should
be chosen in stands of superior provenance.
- 203 -
(ill) Imported seed. This may be from good phenotypes of suitable provenance in
natural stands or excess seed from an external breeding programme using an appropriate
provenance grown in similar conditions. Improved imported seed should only be used to
enrich a locally selected gene pool, it can never sibstitute local selections.
Base for future selection. At an early stage in the breeding programme, it is
appropriate to begin the importation of select genes (as seeds, scions or pollen) to broaden
and improve the genetic base available locally for second generation selection. There are
several examples of success in this procedure: scions, open-pollinated seeds and pollen of
some 85 select trees of Pinus pinaster imported into W. Australia in the mid-sixties;
vigorous exchange of select genes of P. radiata among several countries; scions and seeds
of many P. caribaea var. hondurensis select trees exported from Queensland, Australia, to
enrich gene pools elsewhere; seeds from selected orchard clones and ortets of P. eliottii
and P. taeda in the U.S.A. exported to several countries in the southern hemisphere.
It is necessary to watch closely for diseases and insects in any exchanged material
and especially in scions. Watch for potential undesirable hybridization when planting
important breeding material; interspecific crossing has caused a lot of trouble with
eucalypts in Brazil.
Progressive improvement of seed through successive orchards and clone banks
The above means for providing somewhat improved seed to meet immediate requirements
should be implemented as soon as possible in the programme. However, such techniques give
relatively small gains. Higher gains are achieved in classical seed orchards, although the
time-lag is longer.
(i) Production orchards. A compromise and a useful strategy to gain time, is the
establishment of clonal (or seedling) seed orchards (annually using the best material
available each year.
(ii) Progeny populations. Controlled crossing is used for the production of
families for second-generation selection. Information for roguing the orchards is obtained
from progeny tests established soon after parental selection, as well as from the full-sib
progeny studies in which selection is carried out.
(iii) Breeding population in clone banks; As explained earlier the seed production
population is continually reduced to a superior nucleus (never less than some 25 clones
per production orchard); the roguing is based on progeny trials. Parallel with these
orchards it is essential to create highly variable, high-quality new selection populations
for successive generations of selection. This requires the mating or large numbers of
unrelated trees. It can be accomplished by vegetatively propagating select trees and making
the crosses in a clone bank established for this purpose. Thus, large diverse selection
populations and relatively narrow seed production populations are continuosly generated and
selected. Special measures may be required to maintain diversity in breeding and selection
population such as importation of new material or production of "wide crosses' 1 (Zobel and
Me El wee, 1964; Zobel et al^ 1972). If adequate precautions are taken to keep a "sufficient"
number (>50) of unrelated individuals in the breeding population, one would expect to be
able to make steady progress for many generations without needing to add comparatively
unselected new stock simply to restore lost genetic variability.
Conservation of gene resources. The tree breeder must be concerned with conservation
of the diversity found in the natural populations, both for future research and as a source
in the future of gene combinations not currently required (Kemp et al, 1972; Nikles 1973-b;
Yeatman, 1973; Zobel, 1973). Strategies are discussed in the lecture on conservation,
Demonstrating progress. Libby (1973) listed 'political demonstration' as useful forms
of planting. These are conveniently located plantings of simple design, the purpose of
which is to show the achievements of the programme in as unambiguous a manner as possible.
It may be desirable to deliberately include poor material, e.g. unsatisfactory provinces
as well as 'routine' controls and 'improved' material.
- 204 -
STAFF ORGANIZATION, FACILITIES AND ADMINISTRATION
Staff and facilities; It is essential to appoint an interested, well-trained and
capable professional tree breeder to plan and develop the breeding programme in consulta-
tion with senior managerial and utilization personnel. For a large-scale programme,
the tree breeder should be supported by well-trained technicians and labour of high
quality.
At the headquarters, or within easy access, there should be biometrical services
and library facilities. It may be necessary to take special steps to develop or obtain
efficient techniques and computer programmes designed to facilitate the collection,
processing and analysis of data. The tree breeder and the biometrics team should keep
abreast of modern developments in mating and field designs and work on their application
to local problems. Opportunity should be provided for the tree breeder to participate in
regional, national and international scientific meetings, workshops and refresher courses
from time to time. The local group or team of workers at headquarters should be encouraged
to hold occasional seminars and discussions at which progress and problems being
encountered in their programmes, each contributing to the overall aims of management, are
reviewed.
Cooperation with other tree breeders working with the same species under similar
conditions is most worthwhile, and periodic reciprocal visits by the breeders among
programmes should be arranged.
All these aspects help to ensure constant review of the direction and aims of the
work and a vitality and sense of purpose in the personnel.
Revisions of plans and publication of results; The tree breeder plan should be
revised periodically when thorough re-examination of objectives and approaches is made.
Annual reports and work programmes should be prepared for consultation with administration
officers and review by colleagues or willing experts.
Publication of results should be seen as an obligation to colleagues, as a means of
informing sponsors and administrators of progress (positive and negative results) and
problems, and as an intellectual discipline for the tree-breeding team.
SUMMARY
Planning an efficient tree improvement programme tailored to the requirements of
forest management and utilization and resources available will involve the following main
activities and decisions:
(i) Define appropriate aims for the programme, especially concerning its duration,
the improvements sought and the characteristics that are best manipulated by genetic means.
(ii) Ascertain the resources that will be made available at least for the first
several years. Appoint and encourage an interested tree breeder to help plan and run the
programme with appropriate assistants.
(iii) Locate suitable populations, in terms of provenance, size, age, sites, etc.,
for selection.
(iv) Assemble relevant information on the biology of the species and decide the
areas in which further research is required. Confirm or develop suitable techniques for
grafting, pollinating, raising plants, etc.
(v) Choose appropriate selection criteria, and limit their number,
(vi) Develop methods for rapidly obtaining seeds of better quality.
- 205 -
(vii) Work out long-term dynamic, flexible breeding strategies no that a
sufficiently large number of trees can be established in the breeding population, taking
into consideration the planned programme and the number of generations for which
improvement programmes are planned.
(viii) Work out dynamic long-term and flexible breeding strategies having regard
to aims, resources, time-scale, seed requirements, species characteristics, other
institutes, regions and countries working on the same species should be carefully
considered*
(ix) Set up efficient administrative arrangements and make provision for periodic
reviews of the programme and its modifications as necessary.
*********************
BIBLIOGRAPHY
Allard, R.W., 1960. Principles of Plant Breeding. (Wiley, N.Y.).
* An6n, (1978). Documentos de la Tercera Consulta Mundial Sobre Mejora de Arboles Fores tales.
CSIRO. Canberra, Australia.
Brown, A.G.. 1973. Notes on IUFRO Quantitative Genetics Workshop - Tokyo, 1972. Proc.
Third Meeting Reps., Rec. Working Group No. 1, Mt. Gambler, 1972. App. 17: 1-6. Forestry
and Timber Bureau, Canberra, Australia.
Burdon, R.D., and Shelbourne, C.J.A., 1971. Breeding Populations for Recurrent Selection -
Dilemmas and Possible Solutions. N.Z. J. For. Sci. 1 (2): 174-193.
* Burdon, R.D. and Shelbourne, C.J.A., and Wilcox, M.D. Advance Selection Strategies. Proc
3rd World Consult. For. Tree Breeding. FO: FTB- 77-6/2. Canberra, Australia.
* Burley, J., and Nikles, D.G. (EdsJ Selection and Breeding to Improve Some Tropical Conifers.
I, II. Commonw, For. Inst. Oxford, U.K. (1972/73).
Burley, J., and Nikles, D.G., 1973. (Eds.) Tropical Provenance and Progeny Research and
International Cooperation. Commonw. For. Inst. Oxford, U.K.
Burrows, P.M., 1967. Seed Orchard Systems for Tree Breeding. Rhod. Zamb. Mai. J. Agrio.
Res. 5: 273-280.
Burrows, P.M., 1970. Co ancestry control in forest tree breeding plans. 2nd meeting of
Working Group on Quant. Gen. Eds. Namkoong and Stern. South. For. Exp. Sta. New Orleans.
Co sco, J.N., 1970. Genetic Selection Criteria in Pinus radiata. M.Sc. Thesis'. 122 pp.
Aust. Nat. Univ., Canberra, Australia.
porroan, K.W. , 1976. The Genetics and Breeding of Southern Pines. USDA FS Handbook 471.
Eldridge, K.G., Brown, A.G. and Matheson, A.C., 1977. Genetic gain from a Pinus radiata wood
orchard.
* FAO, 1973. Informaci6n sobre Recursos Genticos Forestales. Doc. Occasional For. 1973/1.
38 pp. (FAO. Rome)
* Faulkner, R. 1975. Seed orchards. Gt. Br. For. Coram. Bull. 54
* Kemp, R.H., BurUy, J., Keiding, H., and Nikles, D.G., 1972. International cooperation in
the exploration, conservation and development of tropical and sub-tropical forest gene
resources. Seventh World Forestry Congress. Argentina. 1972.
Libby, W.J. f 1969. Seedlings vs. vegetative orchards. FAO - North Carolina State Univ. Forest
Tree Improvement Centre, 1969. Lecture Notes: 306-16. School of Forest Resources, N.C.
State Univ., Raleigh.
* Libby, W,J. f 1973. Domestication strategies for forest trees. Can. J. For. Res. 3, 265-76.
- 206 -
* Namkoong, G., 1966. Foundations of Quantitative Forest Genetics* 85 pp. The Govt. For.
Expt. Station of Japan.
Namkoong, G., Snyder, E.B., and Stonecypher, R.W. Heritability and Gain Concepts for
Evaluation Breeding Systems such as Seedling Orchards. Silvae Genetica 19(3): 76-84.
Nikles, D.G., 1970. Breeding for Growth and Yield. In; Forest Tree Breeding. Unasylva 24
(2-3), 97-98: 9-22.
Nikles, D.G., 1973-a. A Proposed Breeding Plan for Improvement of Caribben Pine (Pinus
caribbaea More let var. hondurensis Barr. and Golf.) Based on International Cooperation.
In: Burley, J., and Nikles, D.G. (Eds.), 1973 - Selection and Breeding to improve Some
Tropical Conifers. 2. (Commonw. For. Instit., Oxford, England).
* Nikles, D.G., (1973-b), Biology and Genetic Improvement of Araucaria cunninghamii Alt.
in Queensland, Australia. In; Burley, J., and Nikles, D.G. (Edsj, 1973 - Selection and
Breeding to Improve Some Tropical Conifers. 2. (Commonw. For. Instit., Oxford, England).
* Nikles, D.G., 1974. Planning a tree improvement program. Report on the FAO/DANIDA Training
Course on Forest Tree Improvement. FAO/DEN/TF 7127 (FAO, Rome) 22 5-4 2.
* Nikles, D.G., Burley, J., and Barnes, R.D., 1978. Progress and Problems of Genetic Improvement
of Tropical Forest Trees. I, II. Commonw. For. Inst., Oxford, U.K.
* Porterfield, R.L., 1978. Economic Evaluation of Tree Improvement Programmes. Proc. 3rd
World Consult. For. Tree Breeding. FO: FTB-77-5/2. Canberra.
* Reilly, J.J., and Nikles, D.G., 1978. Benefits and Costs of Tree Improvement; Pinus
caribaea. Proc. 3rd. World Consult. For. Tree Breeding. FO: FTB-77-5/3. Canberra.
Roche, L. (Ed.),* 1978. Metodologia de la Conservaci6n de los Recursos Gene*ticos Forestales.
FO: MISC/7518. FAO. Rome. ^
Shelbourne, C.J.A., 1969. Tree Breeding Methods. Tech. Paper No. 55. 43 pp. New Zealand For.
Serv. , For. Res. Instit., Rotarua.
* Shelbourne, C.J.A., 1971. Planning Breeding Programme for Tropical Conifers. N.S. For Serv.
Reprint No. 548. 20 pp. For. Res. Instit., Rotarua.
Stonecypher, R.W. , 1966. The Loblolly Pine Heritability Study. Internt. Paper. Co. Tech.
Bull. 5:1-128
Stonecypher, R.W. , 1970. Multiple trait breeding. Unasylva 24 (2/3), 48-51.
* Teich, A.M., and Carlisle, A., 1978. Anllisis de Costos y Beneficios en Prograroa de Mejora
Genetica Forestal. Unasylva Vol. 30, No. 118/119.
* van Buijtenen, J.P., 1975. The planning and strategy of seed orchard programs, including
economics. In: Faulkner, R. Seed Orchards, Gt. Br. For. Comm. Bull. 54
van der Meiden, H.A., 1978. Economics of Poplar Breeding. 3rd World Consult. For. Tree
Breeding. FO: FTB-77-5/4. Canberra,
Willan, R.L., and Palmberg, C., 1974. Improved use of forest genetic resources. Report on
the FAO/DANIDA Training Course on Forest Tree Improvement. FAO/DEN/TF 112 (FAO, Rome).
90-101.
Wright, J.W., 1976. Introduction to Forest Genetics. (Academic Press, N.Y.).
Yeatman, C.W., 1973. Gene Conservation in Relation to Forestry Practice. Thirteenth
Meeting of the Committee on Forest Tree Breeding in Canada. Proc. 2: 13-17. Canadian
For. Serv., Ottawa.
* Zobel, B.J., 1973. Gene Preservation by Means of a Tree Improvement Programme. Thirteenth
Meeting of the Committee on Forest Tree Breeding in Canada, 1971. Proc. 2: 13-17.
Canadian For. Serv., Ottawa.
Zobel, B.J. and McElwee, R.L., 1964. Seed Orchards for the Product ion of Genetically Improved
Seed. Silvae Genetica 13 (1-2); 4-lTT
Zobel, B.J., Weir, R.J, and Jett, J.B., 1972. Breeding Methods to Produce Progeny for
Advanced-generation Selection and to Evaluate Parent Trees. Can. J. For. Res. 2; 339-345 .
The references nmked with an asterisk (*) have already been quoted in the tt of the report.
- 207 -
CU
c
cd
o
0)
0)
Q>
CO
o
a
o
0(
c
4 e
H
H|!
b ^ 6 +
(0
u
g ^
K
-
i
ill
irll
1 s
h*
:i
3*3
i&?
s I
O H t
il^J
O C 4* C H J*
!i!;;Ki
' * O 4 H * C
isil^yr
i^llM
Jl
3*2
:ll
m
* i
?=,!
ill"
!Mi
^
"!!
8
6
i
i.i:
:?!
it!.
as-s
|:i!
1
a
eo
Firtt f
cloaal ore
Hi
fl
- 208 -
as
<<
i-a
10
O
95
H
SB
H
U
H
w
s
H
CO
O
as
M
Q
H
H
Ptf
W
O
w
CO
N
N
cd
O
r
O
O
U3
2
-< O
B
CU
s
g
en
g
O
d
ao
c
1-4
S -H
3 00
O JJ
CO
u-i cd
O 1-1
00 x:
C 00 >
^ C J=
c -^ o
3 .* <3 yj
- M - cd n 7
E- -O
5 ^
O en
~j w
O H
si
<- U-
o- en
W V2
z o
o o z
Q ^3 h-4
85 fc
CUi c-i
Cx3 fa tD
2 O O
CiO
c
1-1
c
" M
cu
en
tf
r-l
^5
>,
O
cd
)-
jQ
CO
00
C
3
O
O .
00
H
^
cd
H
H
03
II
CO I
O-,
o
en
s
ii
5
13
ua
en
^J
i
t-i
o
01
4-
cd
M
4J
CO
0)
o
0)
2
GO
01
CO
en
H
UJ
en
en
_
V
>
^H
z
o
tfCrf
o
5
\
PLANTATI
SELECT I
>
>-
r \
=
^
5-
^
^
CJ Z
etf O
1
S 2
O M
^j
o M
H
CO
H
Q ^
S
w
en z
0)
J- C
o
>% *
a <
tj o2
y^^ tjj
CO Z
u
r^^ fj
5
c
4^
cu
X
w o
h Qj
oc
^J en
^
l-t 0>
a- en
O Q
dS
en
- 210 -
Figure 4. Stages in the use o Eucalyptus cuttings
PROGRAMMED ACTION
SELECTION OF A HIGHER
QUALITY PHENOTYPE
Stump coppicing
REJUVENATION BY
COPPICE SHOOTS
Collection of shoot crop
lut tings placed in mist conditions
ICLONAL TRIAL NO
13 F*
|C1
GENERATION
CLONAL BANK
CLONAL TRIAL
NO. 2
OBSERVATIONS
The sapling will have revealed
its potential between 4 and 10
years of age
Well exposed stumps cut in November
3 to 6 rotations on the stump during
the same year
*
Clonal test: 50 saplings spaced at
4 x 4 m
In the clone bank they are spaced
at 5 x 5 m
Fertilization
Coppicing of stumps
at three years old
REJUVENATION BY
COPPICE SHOOTS
Coppicing starts in October
Gathering of shoots
X
Cuttings under mist
_ %l*
{
2 rotations on the stump during
one year
COMPLEMENT OF
CLONAL BANK
IMPROVED MULT ICLONAL
STANDS
2nd GENERATION
CLONE BANK
1st coppicing at 3 years
r The second generation clonal
bank is created depending on
results from clonal trial No. 1
It receives the same treatment
as the first generation clone
bank. One third is coppiced
each year
Coppice rotation 3 years
HIGHLY IMPROVED
MULTICLONAL
STANDS
- 211 -
o
00
o
0)
o
en
0)
a
o
(U
u z
as
sa
C co
iK
U*
O
ss
2S
H (d
g
s
I
JL
38
5
E u
"~ UJ
u. ?
O H
M >-
U CM
UJ
c-
si
(O W
Sd
CO
S3
O
00
H
o
0)
CD
a/
k4
3
8-
Id
9 2
et o
0Q H4
= 1
W M
II
en
Q
pLIiJ
NS
SB
S3
Z
O
is
b.
co Id
ii
H U
ON
^S
dSn
M
2S^
8
o
I
H
- 212 -
6
S
J
3
a
i
- 213 -
SOME ASPECTS OF THE PROBLEM OF GENETIC IMPROVEMENT OF
HARDWOOD SPECIES NATIVE TO VENEZUELA
Marcelino Quljada R.
Institute de Silvicultura
Universidad de Los Andes
Mfirida, Venezuela
OVERALL SITUATION
In Venezuela, as in many Latin American countries, hardwood species have not played
a very important role in plans for genetic improvement, most of which are only in their
initial stages. This is due largely to the fact that the natural forests, which are rich
in flora, particularly broadleaved trees, have so far been expected to meet only local
requirements, one species being replaced by another as time goes on. Another factor is
the relatively low demand, particularly as regards quality, in many tropical regions.
The weakness of the forestry sector within national economies has also had an effect:
the forest enterprises are usually small, and though they use forest resources, they do
not generate them. The larger enterprises have been centred around the coniferous
species, particularly those belonging to the genus Pinus , for pulp and paper making.
The history of the native broadleaved species can therefore be summed up as
follows:
a) Deterioration of the genetic base of species which have been of commercial
importance at one time or another. Typical examples are the species of the Mellaceae
family. This situation is due basically to continued dysgenic selection in the natural
range of these species. This kind of negative selection has well been called trader's
selection since the best trees in the forest have been logged, leaving as a source of
regeneration the trees with the worst phenotypic characteristics.
b) Ignorance of the inherent and cultural properties of the species. This is
due in large part to the feeling that the natural forests are inexhaustible, and to the
problems in the initial utilization of some species. The latter is influenced by the
use of genetically poor basic material and the application of cultural techniques
developed specifically for temperate-zone species. However, the situation has begun to
change in many places, due to the following factors among others:
1) exhaustion of natural sources of wood, particularly near large population
centres ;
2) an increase in demand owing to the growth in population;
3) the need to diversify economies traditionally dependent on a few
resources, particularly non-renewable resources. Here forestry
economics is particularly important ;
4) the emergence of a desire to recuperate natural ecosystems and create
recreation areas.
Some countries have proposed as a solution the use of introduced broad leaved
species, such as Tectona. Gmelina and Eucalyptus, in view of their favourable produc-
tivity characteristics.
- 214 -
Even so, it has become necessary to increase the use of native species, either as
means of recuperating and preserving the autochthonous ecosystems, or because of the fear
that after a time the introduced species may present problems which outweigh the advan-
tages they offer. The main advantage of the native species, in any case, is biological
adaptation they have achieved in certain sites through generations of natural selection.
SPECIFIC CASES
If greater use is to be made of native species in genetic Improvement programmes,
it is essential that a minimum amount of knowledge on these species be available, covering:
a) Natural variation in those areas of the country where the species of interest
are still found, in order to know the morphological, anatomical, physiological and techno-
logical variants that imply hereditary adaptation to various site conditions for selection
purposes.
b) Phenology of the species, in order to assess the seed production capacity for
supplying plantation programmes. Flowering habits also need to be known for the purpose
of controlled pollination.
c) Cultural habits, from the nursery production phase to plantation techniques and
methods and subsequent tending. The purpose is to reduce the Influence of this component
as much as possible, maximizing the genetic potential for use of plantation sites. Of
particular interest in this phase, for the purposes of Improvement, is the species vege-
tative propagation habit.
d) Development habits under cultivation, including quantitative and qualitative
growth and technological properties. This knowledge serves as a basis for estimating
gains in the possible selection systems that may be applied.
Once this knowledge is available, genetic improvement would be directed towards
aspects of a general nature and of a particular nature. The first category, common to all
species, includes the growth phase, which is essential for a greater yield per unit of
area and for enabling native species to compete with Introduced ones - an aspect which
is very commonly emphasized in justifying the introduction of species. In this connec-
tion the programmes would be directed towards the formation of quick-growing subpopula-
tions, through the selection and testing of vegetative propagules or seed progeny.
In the second category account would be taken of particular problems of species or
groups of species which might result, through appropriate crossing, in a more effective
use of the product. In this connection special mention should be made of the following,
which are among Venezuela's most valuable timber species.
a) Bombacopsis quinata. Efforts are being made to control the development of
buttresses, in order to increase the amount of utlllzable bole wood for sawmill Ing.
Since the buttresses, like the thorns, seem to be a character inherently linked with the
evolution of the species, the aim is not to eliminate them completely, but to reduce
their development to a minimum, through a process of selection and crossing specific
individuals, on the basis of observations and preliminary results with forest trees and
progeny.
b) Meliaceae species. Of particular interest are species of the genera Cedrela and
Swietenia, because of their resistance to Hypsipyla. Since in most part of the world
where these species grow naturally they have been subjected to intensive dysgenlc selec-
tion, the genetic base seems to be fairly weak. A possible improvement strategy would be
to effect intra-specles selection followed by inter-species hybridization, with possible
subsequent backcrossing. Crosses between Cedrela odorata, . angustifolia, . fissllis
an< * PQPtana deserve particular attention, and knowledge of the floral biology of these
species should be intensified. Crosses between Swietenia macrophylla, J5. humills and
Are also being considered. Spontaneous inter-specific hybridization in this
genus has already been reported. The possibility of Intergeneric hybrids, e.g. Cedrela x
Toona, is not to be rejected. The problems Involved may be considerable, but the genes
are morphologically so similar that until fairly recently Toona cillata was classified
under the genus Cedrela.
- 215 -
c) Tabebuia rosea. Efforts will be made to correct ramification habits, particu-
larly the typically polydichotomlc bifurcation of the main trunk. This phenomenon
appears to be connected with Inadequate individual sources of seed and high herltablllty
of the characteristic. Strict individual selection, particularly since stands with very
good Individuals are now available, should lead to important achievements in the near '
future.
d) Phitecelobium saman. The attractive feature of this species is its trunk-
crown ratio. In the open, under a sllvo-pastoral system, the tendency is towards the
development of a very broad crown, to the detriment of the bole quality. This tendency
has also been observed in the natural forest, where, however, the trunks are better.
Through a process of continuous individual selection, with interprovenance crossings,
a broad crown with better trunk quality can be sought in the open field, and a smaller
crown and thicker trunk in dense plantations.
- 216 -
ANNEX V
TRIP REPORT
ANNEX V/l
FLOWERING, SEED PRODUCTION AND ARTIFICIAL
POLLINATION OF BOMBACOPSIS QUINATA IN VENEZUELA
Marcelino Quljada R.
Institute de Silviculture
Universidad de Los Andes
Mfirida, Venezuela
INTRODUCTION
Bombacopsis quinata is one of the most important dry tropical forest timber
species in Venezuela. It is found on various kinds of soils ranging from well drained
terrace soils to poorly draining swamp soils.
Rainfall distribution In the Bombacopsis quinata range oscillates between 1 000
and 2 000 mm/yr, with a 3-4 month dry season from December to March. The mean annual
temperature is roughly 27C, with maxima of 32C and minima of 22C. !B. quinata
wood is widely used in construction, interiors, exteriors, furniture, cabinet making,
carpentry, etc. The wood density is roughly 0.40, based in kiln dry weight and
volume in green wood.
. quinata has been planted on a moderate scale to enrich natural forest, parti-
cularly in the forest reserves. Various trials have Indicated open field planting as a
promising technique, assuming proper site selection.
Studies on I*, quinata were begun in 1961 to gain a better understanding of this
species for breeding purposes. Initially, vegetative propagation habits were studied.
Bombacopsis quinata responded so well to these trials that it came to be considered an
easily propagated species. Other studies on the flowering and fruit setting habits
were begun in 1968 with the establishment of seed orchards, including open and guided
pollination.
FLOWERING AND FRUIT SETTING HABITS
Bombacopsis quinata are monoic, with hermaphrodite, generally erect flowers. The
stamens surround a pistil with stigma slightly protruding beyond the anthers.
The flowers, 10-15 cm long, are night-blooming. They usually open after 6 p.m.,
mostly from 8-10 p.m. , when the temperature drops below 25C and the relative humidity
rises above 60 percent. The morning after the flowers open, they drop their floral covers
and stamens. The style remains, even after the fruit has formed. The flowering season
runs from the end of October to March and sometimes April, depending on how long the dry
season lasts. The best flowering periods are January and February, which are usually the
driest months. Observations in seed orchards and natural forests Indicate gteat varia-
bility in flowering. Some trees have a prolonged flowering period which lasts throughout
the season. Some flower early in January and some flower late, from February onwards.
This is very important in planning the lay-out of seed orchards. The initial age of
flowering, seams to indicate early flowering. Trees not yet 10 years old have been
observed to flower. Extreme cases of plants grown from seedlings, flowering after only
six months in the field, have also been observed.
- 217 -
Fruit setting begins in January, with a lapse of 50-60 days from flower opening
and pollination to dehiscence.
The fruit is dehiscent. The seeds are fuzz-covered and therefore easily dispersed
by wind.
Seeds from open pollination in seed orchards average 47 per fruit, ranging from
zero (empty pods) to 140. The average in the natural forest is 30 seeds per fruit, with
heavy insect attacks on fruit and seeds. Viable seeds are usually smooth and fairly
resistant when pressed flat with the hand. Non-viable seeds are usually rough and are
easily flattened with a slight pressure of the fingers.
Colouring ranges from light to dark brown with various degrees of segregation
within the raroets of a clone, which seems to be linked to the source of pollen.
Size is heavily influenced by the mother tree, and is reflected in the number
of seeds /kg: 20 000 to 40 000, averaging 32 000 seeds /kg.
Studies Indicate that the species is highly resistant to self ing, which favours
cross-pollination. The low occurrence of selfing seems to be facilitated by the fact
that as the flower is hermaphrodite, the stigma is receptive at the moment the flower
opens but before it scatters its pollen. Pollen dispersal usually takes place 10-15
minutes after the pollen is exposed to wind. Some system of reproductive incompatibility,
likewise favouring cross-pollination is also conceivable.
The findings of several years of seed germination trials on seeds from over
100 seed orchards and natural forest trees, suggest that _B. quinata can be classed as
a genetically clean species inasmuch as negative traits such as albinism or dwarfism
have not appeared to date.
CONTROLLED POLLINATION
The size and open floral structure of the J3. quinata flower has made it much
easier to artificially pollinate the species.
When the flower reaches normal size, both the bud and the flower are isolated in
a cloth sack to avoid damage from insects looking for nectar in the flowers. For arti-
ficial pollination, breeders allow the flower to open naturally. When it opens, the
stamen is cut away with small scissors, carefully avoiding contact with the stigma of
the pistil. The breeder takes advantage of the fact that at this stage, the pollen is
still not easily dispersed.
Once the plant has been emasculated, the stigma is introduced into the appropriate
pollen source, and completely saturated. The cross is then identified and the pollinated
flower is again covered with a cloth bag, perhaps until the fruit falls off but in any
case until such time as the fruit is clearly formed.
The pollen is collected at emasculation. The stamens are placed in a Petrl dish
and allowed to dry for at least one hour after which the breeder shakes the dish to
release the pollen* The pollen is stored in the same container. It has been stored
in an ordinary refrigerator for up to two weeks, with no great loss of viability, as
long as the Petri dish was kept tightly closed and excessive accumulation of moisture
within the container avoided. Controlled pollination has been 50 percent successful,
with figures of up to 60 percent for cross-pollination and 4 percent for selfing.
Variations have been observed in inter-crossing different clones, with the highest per-
centage of success from inter-crossing clones of different provenances.
- 218 -
Success in cross-pollination has varied greatly: very poor years with rates as
low as 25 percent and very good years with rates as high as 80 percent, apparently
related to environmental determinants affecting floral development.
Seed production under controlled pollination can go as high as 87 seeds per fruit,
which is 1.9 times the yield from open pollination in seed orchards. Here also there
have been extreme variations, ranging from empty pods to 150 seeds per fruit.
No great qualitative differences have been noted in the viability of selected seed
with either method. Germination rates have averaged about 90 percent.
Seed has been maintained in the laboratory in paper envelopes for as much as six
months and remained highly viable. At temperatures of around 5C, the seed can remain
under ground in a closed glass container for more than a year without any visible signs
of deterioration. Seed has been stored for about two years in a 40-80 percent sulphuric
acid environment in a chemical drier, the only precaution being to check that the paper
or cloth packages in which the seed was stored had not been corroded by the acid, and
to change them every six months.
- 219 -
ANNEX V/2
PLANTATION PROGRAMME
LLANOS ORIENTALES. VENEZUELA
Climatic feature I/
Coordinates: 62W, 8<>N
Altitude: 50 m above sea level
Temperature: Mean max. 32oc
Mean min. 2QOC
Mean annual 26oc
Rainfall: 1150-1200 mm/yr, with the main rainy season from June to September and
the shorter rainy season from November to mid-January
Winds: Average wind speed 7.8 km/hr, with occasional gales of 60-80 km/hr
Evaporation: Mean annual values: 2000-2100 mm.
Plantation Programme (Pinus caribaea var. hondurensis; Eucalyptus spp. <1%)
Company Site
Reserved
Planted
Anticipated
area
area
annual
(1978)
planting
(ha)
(ha)
(ha)
1.
State
- CONARE Coloradito
60 000
7 000
5 000
Chaguaramas
100 000
22 000
5 000
Cadupo
805
-
_
Centella
12 000
4 000
10 000
- CVG Uverito
150 000
50 000
7 000
Sub-total
322 805
83 000
2.
Mixed
- Forester
60 000
2 600
10 000
- Sipas
30 000
3 000
1 500
- Guayamure
60 000
7 000
1 500
Sub-total
150 000
12 600
3.
Private
4 000
TOTAL Llanos Orientales 2/
472 805
99 600
I/ "Uverito - un bosque en la Sabana". Divisi6n de Relaciones Pfiblicas, Corporaci6n
Venesolana de Guayana, CVG. Publicaci6n No. 1078-652; Tomo I, Folio 20; October 1978.
I/ For country-wide information, see Appendix VI, "Venezuela 11 .
- 220 -
ANNEX V/3
CVG. VENEZUELA; ACTIVITIES AND HUMAN RESOURCE
FOR THE PLANTATION PROGRAMME
(P. caribaea var. hondurensis. Uverito)
Operation
Human Resources
Total Production
Seedbeds ;
Prepare terraces
Weed
Sow
Irrigate
Prepare containers:
Cut cardboard
Staple containers together
Fill, weigh
Transplant;
Extract seedlings
Move to terraces
Open seed holes
Sow in containers
Field work:
Sprinkler irrigation
Fertilizing
Fungacide treatments
Weeding
Fumigation
Terraces:
Transport containers
Line up containers
Fill containers with sand
Compact sand
1 foreman
8 workmen
1 supervisor
1 foreman
10 cutters
60 helpers
6 fillers
4 foremen
1 helper
6 hoers
32 sowers
6 hole-diggers
1 foreman
14 workers
360 000 plants /day
3 foremen
6 people to line up containers
36 helpers
27 people to fill containers
9 people to pack sand
Planting:
Land preparation:
Demarcation of stands
Soil preparation
Insect control
4 supervisors
8 foremen
64 people to fill boxes
90 planters
20 people to spray insects
24 tractor drivers
7 000-10 000 ha/yr
(area planned for
1980 - 7 000 ha;
5 000 ha in containers,
2 000 ha directly
rooted)
- 221 -
ANNEX V/3 (cont'd)
Operation Human Resources Total Production
Planting: (cont'd)
Plantation:
'J *"*
Fill boxes and trucks ao So
Transport plants a a
Plant
o o
Maintenance ; *> *>
^^ ^ <y QI
M
Open up and weed firebreaks *
Prevent and fight fires o>
Control weeds **
Insect control
Additional personnel,
to staff dining rooms,
infirmary , machine shop,
carpentry shop, etc.
TOTAL HUMAN RESOURCES: 180 workers
- 222 -
ANNEX V/4
CORPORACION VENEZOLANA DE GUAYANA; CALENDAR OF ACTIVITIES
(Plnus carlbaea var. hondurensis. Uverlto)
g
H
4-1
Ou
T
^
3"
1 1
1 1
1 1
3
*"*
1 I
0>
Y"
1 1
j
j i
i i
j
i i
*
I
1
1
1
!
t-i
! ! ! !
i
o-
i i i
1
: i i
1
i
,0
1 !
j
Q)
-^ 1
i
^ i
1
-i.
e
1 1 1
7-
<d
T 1 i !
4.
^
,
i
i
4-1
j
i
8
^ J^
a
^
CO
co a
'8
SPECIFIC OPERATION
Seeding in seedbed
Prepare containers
Prepare terraces
Transplant seedlin
Sprinkler irrigati
Fertilization
Trimnlng roots
Weeding
Insect control
CO
1
4-1
CO
O
00
2 3
CO S
Fill crates and tr
Actual planting
Firebreaks
Weeding
Insect control
CO
M
i
I
jS
*
CO >
4J
CO
g
O
1
848 S
H 4J CO
|
^J
Cu
9 "> S
JJ
9
60 2
i
4J
8
fH
> -H (3
32.5
0.
H
f-4
it
A
O
CO
PL,
H
- 223 -
APPENDIX VI - COUNTRY STATEMENTS
COUNTRY STATEMENTS; ARGENTINA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Aera of country: 2 790 000 km 2
1.2 Location: longitude: 53 - 73W; latitude: 21 - 55 S.
1.3 Population: 26 000 000
1.4 Main climatic and vegetative zones: "Selva Misionera" (Misiones Forest) - "Selva
Tucumano Boliviana" (Bolivian Tucuman Forest) - "Selvas en Galerla" (Riparian
Forests) - "Bosques Subantarticos"(Subantartic Forests) - "Parque Chaqueno" (Chaco
Park) - "Parque Puntano-Pampeano" (Puntano Pampas Park) - "Parque Mesopotamico"
(Mesopotamian Park) - "Estepa Pampeana" (Pampas Steppe) - Estepa Patagonica" (Pata-
gonlan Steppe) - "Monte Occidental" (Western Forest) and "Estepa Punena" (Puno
Steppe) (see Annex I).
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 60 300 000 ha (productive forests: 39 000 000 ha),
2.2 Proportion of land under forest: 20.54%.
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
a) Rational utilization in accordance with forestry plans.
b) Utilization with special emphasis on the most valuable end product.
c) Encouragement of reforestation and improvement of national forests.
d) Encouragement of afforestation, encouraged by development credit.
e) Conservation of protective forests.
2.4 Legislation available to implement policy?
YES - Law No. 13.273 of 1948
2.5 Ownership of forests (out of the 39 000 000 ha of productive forests)
Under State control: 18 150 000 ha (46,16%)
Private ownership: 20 850 000 ha (53.84%)
Community ownership:
No effective control:
2.6 Major forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) :
In 1977: _ _. ,.__
Saw-wood 21 000 000 m 2 Plywood 54 000 t
Veneer logs 58 000 m* Charcoal 278 000 t
Logs for fibreboard 225 000 m3 Fuelwood 1 095 000 t
Logs for sleepers 61 000 t Pulp for paper 703 000 t
p s t8 199 000 t Tanin (extract) 87 000 t
2.7 Staff
Professional
Sub-professional
(with diploma or
certificate of training)
2.8 Gross annual budget for forestry: US$ 47 680 000 21.
I/ Only under the National State. .
|/ Only under the National State, including incentive subsidies (US$ 1.00 - $a 1.630).
- 224 -
3.
AFFORESTATION AND REFORESTATION
3.1
Areas
3.1.1
Net total area of plantations at the end of 1978: 530
000 ha - 1 .
3.1.2
Planned annual target of afforestation/reforestation: 90
000 ha/year (1979) -'
3.2
Organization and administration of planting schemes:
3.2.1
State forest services: 100%
3.3
Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.).
3.3.1
Indigenous species
Mean annual increment
(without bark) at the
Net area Rotation
end of the rotation
Product/Purpose Species (ha) i/ (years)
(m 3 /ha/year)
Saw-wood Araucarla -/
17 (man-made planta-
Pulpwood augustifolia 30 000^' 40
tions
Tanin and Schinopsis
6 for the stand, 0.5
sleepers sp. 12 000 000 120
for the species
Saw-wood Nothofagus
3-4
sp. 1 000 000 90
(homogeneous forest)
Saw-wood ( Cedrela sp. )
0.5
( Cordia sp. ) 2 500 000 60
0.15 (for the stand)
( Half our odendron
. 3 (heterogeneous
( sp. )
forest)
3.3.2
Introduced species
Mean annual* Increment
(without bark) at the
Net area Rotation
end of the rotation
Product /Purpose Species (ha) i/ (years)
(m 3 /ha /year)
Paper pulp Pinus elliottii) ,- n nftn 10-25
30
and resin Pinus taeda ) " u uuu 10-25
30
Wood and pulp Pseudotsuga
menziesil 32
20
Pinus radlata 30
25
Pinus ponderosa 35
18
Containers pulp Populus sp. 170 000 10
22
Eucalyptus sp. 100 000 20
30
4.
SEED AND PLANT SUPPLY
4.1
Service and control of seed supply
4.1.1
Have seed zones been defined for Indigenous species (I.e.
, is it common to
restrict the use of seed for afforestation/reforestation
to the particular
zone in which it was collected)?
YES (only with Araucaria angustifolia in Mlsiones)
4.1.2
Is there a national seed certification system? NO 4/
4.1.3
Are there facilities for storing seed at controlled temperatures? YES
4.1.4
Does the supply of seed cover the demand for the species
listed
In 3.3.1? NO in 3.3.2? NO
T7 The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
2J Approximate dates.
3J Corresponds to range of the species in Argentina.
\l In process; status: preliminary draft law.
- 225 -
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? YES I/
5.1.1 If yes, list the species concerned:
Pinus elliottii Populus sp.
Pseudotsuga menziesii Salix sp.
Pinus ponderosa Eucalyptus saligna/grandls
Araucaria angustifolia Melia azedarach
5.1.2 Most important characters to be improved or bred for:
1) Productivity
2) Ecological adaptability
3) Disease resistance
4) Seed quality
5) Wood quality
5.2 Brief outline of improvement methods already applied.
5.2.1 Species /provenance trials (indicate in parentheses the number of provenances
being tested):
a) Pinus elliottii and P. taeda (38 and 29 provenances) 27
b) Eucalyptus camaldulensis, E. viminalis (33 and 22 provenances) 2/
c) Pinus in northwest Argentina (19 species) 2/
d) Pinus on the Argentine coast (17 species) 2J
5.2.2 Area of seed stands in each of the main species: _3/
Pinus ponderosa 8 ha
Pseudotsuga mensiesii 15 ha
Araucaria angustifolia 414.5 ha
5.2.3 Plus trees of the main species (indicate in parentheses the number of trees): 3f
Pinus ponderosa (24)
Pseudotsuga menziesii (17)
5.2.4 Seed orchards (species, area and number of clones or mother trees): 31 NO
5.2.5 Progeny testing? (species and area): 31
Pinus ponderosa (1979)
Pseudotsuga menziesii (1979)
5.2.6 Other methods of improvement (specify): J3/ NO
6. MOST SUCCESSFUL METHODS OF VEGETATIVE PROPAGATION FOR MAIN SPECIES
Species Method % of success
Pinus elliottii Stakes (horm. treatment and Started on 12.12.79
treatment of mother tree)
Eucalyptus saligna " Started on 12.12.79
Pseudotsuga menziesii " Starts on 1.2.80
Pinus ponderosa " Starts on 1.2.80
7. REFERENCES TO TREE IMPROVEMENT
Plfease list the references to tree improvement in the country, in publications,
reports, etc.
IDIA (Suplemento Forestal No. 6)- INTA - Bs.As. 1970
IDIA (Suplemento Forestal No. 8)- INTA - Bs.As. 1973
IDIA (Suplemento Forestal No. 5)- INTA - Bs.As. 1966
IDIA (Revista No. 12) - Buenos Aires
T/ By 1980 at national level.
2/ Institute Nacional de Tecnologla Agropecuaria (INTA) - Natl. Institute of Agricultural
Technology.
31 Only under the jurisdiction of the National Forestry Institute.
- 226 -
PHYTOGEOGRAPHICAL REGIONS OF ARGENTINA
MAP OF THE
REPUBLIC OF ARGENTINA
Phytogeographical regions
CNAOO PARK EASTERN ZONE
CHACO PARK WESTERN ZONE
TVCUNAW - 10LXVIAN FOREST
GALLERY FORESTS
MIStONES FOREST
MESOPOTAMIA FOREST
PUNTANO - PAMPAS PARK
WESTERN FOREST
ANDEAN DESERT
PAMPAS STEPPES
PATAOONIAN STEPPES
Sub-in! rrrir fart. northern
or
Sub-ant *rrtir for*t 4.i<HtHern
Argtnttn* Antiirctir f
South Ortn+r ^ u
- 227 -
COUNTRY STATEMENT; BOLIVIA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 1 058 000 km 2
1.2 Location: longitude: 5726 f - 6038'W; latitude: 938 f - 22O53 f N
1.3 Population: 4 647 816 (September 1976)
1.4 Main climatic and vegetative zones: mountain-building in the area has produced
four different zones: warm valleys, valleys, high plateaux and plains.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 56 468 000 ha
2.2 Proportion of land under forest: 51%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
a) To formulate the country's forest policy and its plans for implementation.
b) To manage national forest resources on a permanent basis.
c) To promote and/or make a forest Inventory for Bolivia.
d) To authorize, guide and supervise forest utilization, in accordance with
the provisions of the law.
2.4 Legislation available to Implement policy? YES
2.5 Ownership of forests:
Under State control: 99.9 percent
Community ownership: 0.1 percent
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) :
Veneer logs, sawn-logs, sleepers, posts, planting for mines, building timber,
charcoal fuelwood, rubber and chestnut.
2.7 Staff State Other
Professional 28 5
Sub-professional (with diploma or certificate of training) 21 5
2.8 Gross annual budget for forestry: US$ 1 000 000.
3. AFFORESTATION AND REFORESTATION
3*1 Areas
3.1.1 Net total area JL/ of plantations at the end of 1978: 10 000 ha.
3.1.2 Planned annual target of reforestation: 1 500 ha/year.
3.2 Organization and administration of planting schemes:
3.2.1 State Forest Services: 90 percent
C^MARAFOR: 3 percent
Other private owners: 7 percent
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and
stakes, pulpwood, fuelwood, protection, etc.).
I/ The net area Is the gross area of plantations, less the area occupied by roads,
paths, firebreaks, buildings and other areas not covered by trees.
. 228 -
3.3.1 Indigenous species
Product /Purpose
Saw-wood
Species
Swietenia
macrophylla
Net area
(ha) !/
Rotation
(years)
Mean annual Increment
(without bark) at the
end of the rotation
(m3/ha/year)
200
80
Sulphate
"Quina"
42
10
Latex
(rubber)
Hevea
brasiliensis
230
10-20
Fuelvood
3.3.2 Introduced species
Product /Purpose
Posts
Charcoal
Stakes
Posts
Others
Species
Eucalyptus spp
228
Met area
(ha) I/
9 000
300
Rotation
(years)
10
20
Mean annual increment
(without bark) at the
end of the rotation
(m3 /ha/year)
20
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? YES
4.1.2 Is there a national seed certification system? NO
4.1.3 Are there facilities for storing seed at controlled temperatures? NO
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1? NO
in 3.3.2? NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement progr ?
5.2.5 Progeny or testing? (species and area):
Eucalyptus rostrata Pinus radlata
E. vimlnalls P. elllottli
E. saligna P. taeda
E. Globulus Cupressus lusitanica
E. citriodora
E. teretlcornis
- 229 -
COUNTRY STATEMENT; BRAZIL
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 8 511 970 km 2
1.2 Location: longitude: 34S-74W; latitude: 5N-33S
1.3 Population: 119 670 000
1.4 Main climatic and vegetative zones: See Annex I
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 352 000 000 ha
2.2 Proportion of land under forest: 41%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
a) Forest production to meet the National Programmes for pulp and paper and
for the charcoal fueled steel industry
b) Reforestation Programme for small and medium-sized rural holdings
c) Settlement and rational utilization of the forest resources of the
Brazilian Amazon region
2.4 Legislation available to Implement policy? YES
2.5 Ownership of forests
Under State control: 75%
Private ownership: 15%
Community ownership:
No effective control: 10%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) :
Saw-wood 16 325 000 m 3 Charcoal 17 618 000 t
Sawn wood 7 400 000 m3 Paper 2 045 969 t
Pulp 1 253 784 t Veneer 695 000 m 3
Fibre board 574 000 m 3 Particle board 461 000 m3
Plywood 659 000 m 3 Fuelwood 140 000 000 m 3
2.7 Staff State Others
Professional 1 200
2.8 Gross annual budget for forestry: US$ 30 266 343.83
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978: 3 319 034 ha
3.1.2 Planned annual target reforestation: 335 442 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State forest services: 5%
Others: 95% (most of this due to tax incentives)
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and
stakes, pulpwood, fuelwood, protection, etc.).
I/ The net area Is the gross area of plantations, less the area occupied by roads,
paths, firebreaks, buildings and other areas not covered by trees.
- 230 -
3.3.1 Indigenous species
Product/Purpose Species
Araucaria
angustifolia
Indigenous
species
3.3.2 Introduced species
Product /Purpose Species
Pinus spp.
Net area Rotation
(ha) II (years)
73 074.91
39 744.75
40
40
Mean annual increment
(without bark) at the
end of the rotation
(m3/ha/year)
13
Net area
(ha) !/
Rotation
(years)
Mean annual increment
(without bark) at the
end of the rotation
(m3/ha/year)
.05 533.82
'39 781.35
7
7
19
20
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e., is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? YES
4.1.2 Do special seed production areas exist? YES
4.1.3 Are there facilities for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1? YES
in 3.3.2? NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? YES
5.1.1 If yes, list species involved:
Pinus caribaea var. hondurensis
P. caribaea var. caribaea
P. elliottii var. elliottii
Araucaria angustifolia P. oocarpa
5.1.2 Most important characters to be improved or bred for:
Eucalyptus grandis
E. urophylla
E. saligna
Volume (dbh and height)
Form
Wood quality
Resistance to diseases
5.2 Brief outline of improvement methods already applied:
5.2.1 Species /provenance trials (indicate in parentheses the number of provenances
being tested):
Eucalyptus spp.
Pinus caribaea var. hondurensis
P. caribaea var. caribaea
P. caribaea var. bahamensis
P. oocarpa
43 species, 1-37 provenances of each species
(7) I/
(2) I/
(2) I/
(19) I/
II
Provenances introduced by PRODEPEF between 1971 and 1977.
- 231 -
5.2.2 Seed stands: area of seed stands In each of the main species:
Eucalyptus grandis 114.99 ha
E. urophylla 36.46 ha
E. saligna 2.30 ha
E, citriodora 8.15 ha
E. paniculata 10.45 ha
E, vimlnalis 215.28 ha
Pinus oocarpa 1 069.58 ha
P. carlbaea var. hondurensis 1 202.96 ha
P. elllottli var. elliottii 30.90 ha
P. taeda 106.50 ha
P. carlbaea var. carlbaea 346.16 ha
P. caribaea var. bahamensls 185.48 ha
P. kesiya 117.87 ha
P. strobus var. chap ens is 6.98 ha
5.2.3 Plus trees of main species (Indicate in brackets the number of trees):
Eucalyptus grandis (200) P. caribaea var. hondurensis (300)
E. saligna (50) P. carlbaea var. bahamensis (300)
Pinus oocarpa (100) P. caribaea var. caribaea (300)
P. taeda (50)
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Eucalyptus grandis (100 clones)
E. saligna (50 clones)
Pinus caribaea (300 clones)
5.2.5 Progeny testing?
Eucalyptus grandis (15 tests) P. caribaea var. hondurensis (22 tests)
E. saligna (5 tests) P. kesiya (10 tests)
5.2.6 Other methods of improvement (specify):
Area of seed collection
Grafted seed orchard
Obtaining and use of hybrids
6. MOST SUCCESSFUL METHODS OF VEGETATIVE PROPAGATION FOR MAIN SPECIES
Species Method % of success
Eucalyptus camaldulensis Estaqula 60
E. tereticornis Estaquia 50
E. grandis Enxertia 60
E. dunnil Enxertia 70
E. dunnii Estaqula 50
P. kesiya Enxertia 80
P. caribaea Enxertia 70
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree Improvement in the country in publications,
reports, etc.
IPEF Revista IUFRO Proceedings
SIF Boletin Tgcnico Revista Arvore
IPEF Circular Tficnica PRODEPEF Sgrle Tgcnica
Silvlcultura en Sao Paulo
ANNEX 1
- 232 -
COUNTRY STATEMENT: BRAZIL
ZONEAMENTO BIOCLIMATICO PAR)
REFLORESTAMENTO
Hqiics iiociiyiTiui tiw. * * ***
- 233 -
Country statement: CHILE
22 2
1.1 Area of country: 2 007 000 km - 757 000 km mainland and islands, and 1 250 000 km
1. GENERAL GEOGRAPHICAL INFORMATION
Area of coi
Antarctic.
1.2 Location: longitude: 6730' - 7530'W; latitude: 17 - 56S (mainland). However,
it stretches down to the Antarctic.
1.3 Population: 11 000 000
1.4 Main climatic and vegetative zones:
Climatic a) Sub-tropical arid
b) Warm temperate with adequate humidity
c) Rainy temperate
d) Tundra
e) Cold steppe
f) Freezing climates
Vegetative a) Xeromorphic zone
b) Mesomorphic zone
c) Hygromorphic zone
d) Andean zone
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 20 000 000 ha
2.2 Proportion of land under forest: 26 percent - taking into account both mainland
and island Chile.
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it? .
To achieve accelerated growth and development of the forest sector through conserva-
tion, protection, management and utilization of forest resources, bearing in mind
the nation 1 s interests and the country's economic and ecological limitations.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
2.5.1 Under State Control: 60 percent
2.5.2 Private ownership: 40 percent
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer
logs, logs for sleepers, posts, piles and stakes, pulpwood):
Chemical pulp, mechanical pulp, newsprint, other papers, cardboard, fibreboard,
particle board, plywood, veneers, logs, sawn timber, posts, stakes, sleepers,
honey, bark, leaves, mushrooms.
2.7 Staff State Others
2.7.1 Professional: 60 300
2.7.2 Subprofessional (with diploma or
certificate of training): 100 360
2.8 Gross annual budget for forestry: US$ 20 000 000.
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978:
* 712 000 ha
3.1.2 Planned annual target/reforestation: 70 000 ha/year
y The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 234 -
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: -
3.2.2 Others (specify):
Small private owners: 45 percent
Large companies: 55 percent
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuel wood, protection, etc).
3.3.1 Indigenous species:
Mean annual increment
(without bark) at the
Product/ Net area Rotation end of the rotation
Purpose Species (ha) I/ (years) (m /ha/year)
Sawn wood Nothofagus spp. 500 annual 80 14
2 000 total
Forage Atriplex 400 annual
repanda 2 000 total
Bark Quillaja 100 annual 40 7
saponaria BOO total
3.3.2
Introduced species:
Mean annual increment
(without bark) at the
Product/ Net area Rotation end of the rotation
Purpose Species (ha) I/ (years) (m 3 /ha/year)
Paper, sawn wood, Pinus 62 000 annual
^>ulp radiata 640 000 total 20-25 22
Firewood, posts, Eucalyptus 3 500 annual
parquet globulus 40 000 total 15-20 25-30
Matches, cases, Populus 2 200 annual
sawn wood spp. 22 000 total 15-20 25-30
Sawn wood Pseudotsuga 700 annual
menziesii 2 000 total 80 13
Forage Atriplex 600 annual
nummularia 3 000 total
4.
SEED AND PLANT SUPPLY
4.1
Service and control of seed supply
4.1.1
Have seed zones been defined for indigenous species (i.e. is it common to
the use of seed for afforestation/reforestation to the particular zone in
was collected)? For very few species.
restrict
which it
4.1.2
Is there a national seed certification system? YES
4.1.3
Are there facilities for storing seed at controlled temperatures? YES
4.1.4
Does the supply of seed cover the demand for the species listed:
in 3.3.1? YES
in 3.3.2? YES
5.
TREE IMPROVEMENT
5.1
Does the country have an official tree improvement programme? YES
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 235 -
5*1.1 If yes, list species involved:
- Pinus radiata, through the establishment of seed orchards.
- Various species involved in the provenance trials (see item 5.2.1).
- Nothofagus alpina, an indigenous species, through a variability study.
5.1.2 Most important characters to be improved or bred for:
Growth rate Volume
General form Wood density
General adaptability Resistance to Dothistroma pini
5.2 Brief outline of improvement methods already applied:
5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being
tested):
Eucalyptus delegatensis (12) Pinus radiata (14)
E. regnans (6) P. ponderosa (6)
E. obliqua (4) P. contorta (3)
E. bicostata (4) P. muricata (2)
E. camaldulensis (3) Pseudotsuga menziesii (2)
5.2.2 Area of seed stands in each of the main species:
Pinus radiata: 120 ha
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
350 plus trees, with a selection intensity of 1 in 120 000.
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Only pinus radiata. There are two types: (1) clonal ; There are 8 of these,
with 42 clones each and areas fluctuating between 11 and 30 hectares, 20 being
the average; (2) from seedlings; there are 3 of these, with 200 families as a
basis, 20 hectares each.
5.2.5 Progeny testing? (species and area):
In May 1980, 55 hectares on 14 different sites, will be put under progeny trials
of Pinus radiata.
5.2.6 Other methods of improvement (specify):
38 000 seed trees of Pinus radiata have been selected, which have produced 7 000 kg
of seed, with an expected improvement of roughly 5-6 percent.
6. SUCCESSFUL METHODS IN VEGETATIVE PROPAGATION FOR MAIN SPECIES:
Species Method 7. of success
Pinus radiata Lateral grafting 80
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree improvement in the country, in publications,
reports, etc.
Burdon, R. Mejoramiento genetico forestal en Chile. Documento de trabajo No. 11.
FO:DP/CHI/76/003 Investigacion y desarrollo forestal. Santiago de Chile, 1978.
Delmastro, R. Proposicion de un programa cooperative de mejoramiento genetico entre
la Universidad Austral, instituciones y empresas forestales. Facultad de Ingenierfa
Forestal, Universidad Austral de Chile, Valdivia. 1976.
Delmastro, R. Informe de genetica forestal. Proyecto CONAF/PNUD/FAO/CHI/ 76/003.
Codigo III - 1.1 Facultad de Ingenierfa Forestal, Universidad Austral de Chile,
. Valdivia, 1977.
Moreno, D.G. Consideraciones preliminares para un programa de mejoramiento genitico.
(To be published in May 1980 in the Supplement to "Chile Forestal").
Moreno, D.G. and Stutz, C. Informe de la gira a Nueva Zelandia sobre mejoramiento
genetico. Documento interno presentado a la Direcci6n Ejecutiva de CONAF a la
Direcci6n del Proyecto CONAF/PNUD/FAO/CHI/ 76/003. Chilian, Chile, 1979.
- 216 -
Delmastro, R. Primer inform anual. Convenio de mejoramiento genltico UACH -
entpresas forestales. Facultad de Ingenieria Forestal, Universidad Austral de
Chile, Valdivia, 1977.
Delmastro, R. Segundo informe anual* Convenio de mejoramiento genetico UACH -
empresas forestales. Facultad de Ingenieria Forestal, Universidad Austral de Chile/
Valdivia, 1978.
Delmastro, R. Manual para los ensayos de progenie de polinizaci6n abierta. Convenio
de mejoramiento genetico UACH - empresas forestales. Facultad de Ingenierfa Forestal,
Universidad Austral de Chile, Valdivia, 1979.
Smith, V.N. Selecci6n de irboles plus en Constituci6n, provincia de Maule. Actas V
jornadas forestales. Asociaci6n Chilena de Ingenieros Forestales. Los Angeles,
Chile, 1969.
Aparico, J. and Siri, A. Estudio de variacion en plintulas de Pinus radiata. D. Don.
en Chile. Tesis Facultad de Agrononrfa, Escuela de Ingenierfa Forestal, Universidad
de Chile, Santiago, 1970.
Garrido, F. Ibarra, M. Steinmetz, J. and Seron, J. Variacion de poblaciones naturales
de RauH (Nothofagus alpina (Poepp. et Endl.) Oerst). Revisi6n bibliograf ica.
CONAF-FAO. FO DP/CHI/ 76/003 . Documento de trabajo No. 28, Santiago de Chile, 1979.
- 237 -
Country statement: COLOMBIA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 1 138 914 km 2
1.2 Location: longitude 7404'41'; latitude 435'56"
1.3 Population: 28 000 000
1.4 Main climatic and vegetative zones:
There are various areas in the country; they include humid and very humid tropical
forest on the Pacific coast and in Amazonia; dry tropical forest on the Atlantic
coast and in the east of the country; and premontane and montane forests in the
Andean area.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 36 426 000 ha (1977)
2.2 Proportion of land under forest: 38 percent
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
To rationally manage forest areas, which are classed as productive, productive-
protective and protective. To encourage watershed management and conservation in
the Andean region. To develop, through support services, such aspects as:
forestation, protection, technical assistance, research and marketing of products.
2.4 Legislation available to implement policy? YES
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Pulp, paper, cardboard - sawn wood - veneers, triplex, particle board, fibreboard -
roundwood - posts, poles, fencing, cross-pieces, charcoal, rubber, balata, balsam of
Tolu, essential oils, gum, etc.
2.7 Staff State Others
2.7.1 Professional: 270 650
2.7.2 Subprofessional (with diploma
or certificate of training): 30 90
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978:
80 000 ha
3.1.2 Planned annual target/reforestation:
40 000 ha /year
3.2 Organization and administration of plantig schemes
3.2.1 State Forest Services: 20 percent
3.2.2 Others (specify):
Forest corporations: 30 percent
Private companies: 50 percent
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.)
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 238 -
3.3.1 Indigenous species:
Product/ Net Area
Purpose Species (ha) I/
Rotation
years
3.3.2
Cabinet- Tabebuia
making rosea 250
Cabinet* Alnus
making jorullensis 500
Cabinet- Cordia
making alliodora 650
Interiors
Cabinet- Cariniana
making pyriformis 700
Introduced species:
30
30
25-30
30
Mean annual increment (with-
Product/ Net area
Rotation out bark) at the end of the
Purpose Species (ha) I/
years rotation (mVha/year)
Pulp Pinus petula 25 000
15-20 20m 3 /ha/year
Roundwood Eucalyptus
7
globuius 18 000
8-15 22m J /ha/year
Pulp Cupressus
3
lusitanica 21 000
15-25 18tn /ha/year
Pulp Pinus
3
radiata 3 000
15 15m J /ha/year
Construction Tectona
3
grand is 1 000
40 15m /ha/year
Round wood Eucalyptus
3
spp. 4 000
12-15 ISin/ha/year
Other species 6 000
various
4.
SEED SUPPLY AND CONTROL
4.1
Service and control of seed supply
4.1.1
Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? YES
4.1.2
Is there a national seed certification system? YES
4.1.3
Are there facilities for storing seed at
controlled temperatures? YES
4.1.4
Does the supply of seed cover the demand
for the species listed in 3.3.1? NO
3.3.2? NO
5.
TREE IMPROVEMENT
5.1
Does the country have an official tree improvement programme? YES
5.1.1
If yes, list species involved:
Pinus patula
Cupressus lusitanica
Pinus oocarpa
Pinus caribaea
Cordia alliodora
Pinus kesiya
Eucalyptus globuius
Eucalyptus grandis
Gmelina arborea
Cariniana pyriformis
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
~ firebreaks, buildings and other areas not covered by trees.
- 239 -
5.1.2 Most important characters to be improved or bred for:
Upward growth Form of the trunk
Specific weight Angle and thickness of branches
Fibre length Crown size
Ease of pruning
5.2 Brief outline of improvement methods already applied:
5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being
tested):
Cupressus lusitanica (20) Eucalyptus globulus (32)
Pinus caribaea (14) Gmelina arborea (11)
Pinus oocarpa (19) Cariniana pyriformis (4)
Pinus kesiya (25) Tectona grandis (5)
Pinus patula (?) Cordia alliodora (13)
Ochroma lagopus (5)
5.2.2 Area of seed stands in each of the main species:
Cordia alliodora
Ochroma lagopus
Jacaranda copaia
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
Cupressus lusitanica (43) pre-selected species
Pinus patula (20)
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Cupressus lusitanica (5 ha, 43 clones)
5.2.5 Progeny testing? (species and area):
Cupressus lusitanica
Pinus patula
Cariniana pyriformis
5.2.6 Other methods of improvement (specify):
SUCCESSFUL METHODS IN VEGETATIVE PROPAGATION FOR MAIN SPECIES:
Species Method
Bombacopsis quinata Large cutting
Tabebuia rosea Large cutting
Cupressus lusitanica Lateral grafting
Pinus patula Grafting
REFERENCES TO TREE IMPROVEMENT
Logros geneticos con Cupressus lusitanica a t raves de 6 afios de mejoramiento geneticc
en Colombia. Carton de Colombia: Investigacion Forestal.
Documentos de Trabajo del Proyecto sobre Investigaciones y Desarrollo Industrial
Forestaies, INDERENA/PNUD/FAO/CONIF COL/74/005:
Konig, A, & Melchior, G.H. Propagacion vegetativa en arboles forestaies;
Konig, A. & Melchior, G.H. & Venegas T.L. Ensayos de procedencias con Pinus
caribaea;
Konig, A. Ensayos de procedencia con Pinus oocarpa y Pinus kesiya en Colombia;
Melchior, G.H. Bases y posibilidades del mejoramiento de arboles forestaies en
Colombia.
- 240 -
Country statement: COSTA RICA
1. GENERAL GEOGRAPHICAL INFORMATION
1*1 Area of country: 50 851 km 2
1.2 Location: Longitude: 8234 f - 8559'; latitude: 802 f -
1.3 Population: 2 125 620
1.4 Main climatic and vegetative zones:
Climatic zones: Valle Intermontano (Inter-montane valley), Vertiente Pacifica
(Pacific slope) , Vertiente Atlantica (Atlantic slope) and the
north.
Vegetative zones: Tropical forest (dry-humid), sub-tropical forest (humid-very
humid) , low montane forest (humid-very humid- rainy) , montane
forest, sub-alpine forest.
2. FOREST AND NATIONAL FOREST POLICY
2.1 Area of forests: 2 088 200 ha
2.2 Proportion of land under forest: 40.9%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
A. To transform productive forests into an efficient system for the supply of
raw material to industry.
B. To conserve protective forests.
C. To encourage large-scale reforestation as a better land use measure. To
strengthen the State institutions for forest planning and control.
2.4 Legislation available to Implement policy? YES
2.5 Ownership of forests:
Under State control: 11.5%
Private ownership: 35.0%
Community ownership: 0.3%
No effective contro:
2.6 Principal forest products (for exemple, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Sixty-five percent of the country's forest production is saw-wood to meet internal
needs, and some for furniture for export. The remainder is processed as plywood
and particle board for export. A low percentage is used for the preparation of
sleepers and souvenirs.
2.7 Staff State
2.7.1 Professional: 20
2.7.2 Subprofessional (with diploma or
certificate of training): 127 29
2.8 Gross annual budget for forestry: US$ 2 264 655.40
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area - of plantations at the end of 1978: 810 ha
3.1.2 Planned annual target /reforestation: 6 200 ha/year
II The net area is the gross area of plantation, less the area occupied by roads* paths,
firebreaks, buildings and other areas not covered by trees.
- 241 -
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 82
Others (Specify): Private ownership: 92%
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.)
3.3.1 Indigenous species:
4.
4,1
4.1.1
4.1.2
4.1.3
4.1.4
Product
Species
Saw-wood Cordla
alliodora
Saw-wood Bombacopsis
quinata
Saw-wood Alnus
acuminata
3.3.2 Introduced species:
Net area
(ha) I/.
Rotation
(year)
20
25
50
25
50
18
Product
Species
Saw-wood Pinus
caribaea
Posts Eucalyptus
deglupta
Saw-wood Cupressus
lusitanica
Pulp
Gmelina
arborea
Saw-wood Tectona
grandis
SEED AND PLANT SUPPLY
Net area
(ha) I/
Rotation
(year)
200
8
50
5-6
100
25
20
25
300
40
Mean annual increment
(without bark) at the
end of the rotation
(m3/ha/year)
18-20
15
35
Mean annual Increment
(without bark) at the
end of the rotation
(m3/ha/year)
30
30
20-22
25
10
Service and control of seed supply
Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? YES
Is there a national seed certification system? YES
Are there facilities for storing seed at controlled temperature? YES
Does the supply of seed cover the demand for the species listed
in 3.3.1? NO
in 3.3.2? NO
? NO
5. TREE IMPROVEMENT
5.1 Does / the country have an official tree improvement progra
7. REFERENCES TO TREE IMPROVEMENT
Please list the reference to tree Improvement in the country, in publications,
report, etc.
There is no literature on this specific subject. The only publications are a few
short information booklets on identification of seed trees, published by the
Forest Service of Costa Rica.
I/ The net area is the gross area of plantations, less the area occupied by road*, paths,
firebreaks, buildings and other areas not overed by trees.
- 242 -
Country Statement: CUBA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 110 922 km 2
1.2 Location: longitude: 74 and 85; latitude: 19 and 24
1.3 Population: 9 405 000
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 1 691 131 ha
2.2 Proportion of land under forest: 15t
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
Reforestation of all potential forest areas to cover an estimated 27 percent
of the country.
Well-managed forests for the purposes of production, protection of the environment
and wildlife, and recreation.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests: Under State control: 100%
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? YES
4.1.2 Is there a national seed certification system? YES
4.1.3 Are there facilitaties for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1? YES
in 3.3.2? YES
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? YES
5.1.1 If yes, list species involved:
Pinus caribaea var. caribaea Hibiscus elatus
P. tropicalis Cedrela odor at a
P. cubensis
P. maestrensis
5.1.2 Most important characters to be improved or bred for:
Vigour
Stem straightness
Remification
5.2 Brief ouline of improvement methods already applied:
5.2.1 Species /provenance trials (indicate in parantheses the number of provenances being
tested) :
Pinus caribea var. caribea (13) Hibiscus elatus (23)
P. tropicalis (6) Cedrela odorata (13)
P. cubensis (10) Tectona grandis (16)
P. maestrensis (14) Eucalyptus saligna (16)
Svietenia macrophylla (8) P. caribaea var. hondurensis (9)
- 243 -
5.2.2 Area of seed in each of the main species:
Pinus caribaea var. caribaea (1 280 ha) P. maestrensis (147 ha)
P. tropicalis (175 ha) Hibiscus elatus (130 ha)
P. cubensis (572 ha) Cedrela odorata (19 ha)
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
P. caribaea var. caribaea (270) P. maestrensis (72)
P. tropicalis (120) Hibiscus elatus (90)
P. cubensis (150) Cedrela odorata (220)
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Pinus caribaea var. caribaea (200 ha) (108 cl.)
P. cubensis (45 ha) (98 cl.)
Hibiscus elatus (13 ha) ,(48 cl.)
5.2.5 Progeny testing? (species and area):
Pinus caribaea var. caribaea (27,5 ha) P. maestrensis (7 ha)
P. tropicalis (21,5 ha) Hibiscus elatus (3 ha)
P. cubensis (6 ha) Cedrela odorata (15 ha)
6. SUCCESSFUL METHODS IN VEGETATIVE PROPAGATION FOR MAIN SPECIES:
Species Method % of success
Pinus caribaea var. caribaea Grafting 80
P. cubensis " 65
P. maestrensis " 70
Hibiscus elatus " 90
Cedrela odorata " 85
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree improvement in the country, in publications,
reports, etc.
7.1 Betancourt, A. and Gonzales A. (1972) "Trabajos realizados en Cuba sobre mejora-
miento gentico de P. carbaea Mor. var. caribaea". In: Special Papers from Cuba
at the Seventh World Forestry Congress, p. 7-26.
7.2 Gonzales, A. (1974) "Algunos aspectos sobre mejoramiento selective de Srboles"
(mimeographed) GIF, La Habana, 19 pp.
7.3 Thomsen, J. (1974) "Gentica forestal y mane jo de semillas fores tales (ditto)
GIF. La Habana, 246 pp.
7.4 Zayas, A. and Bur ley, J. (1973) "Estudio de progenies de P. caribaea var. caribaea"
Cajalbana, in Cuba, 7 pp.
- 244 -
Country Statement: DOMINICAN REPUBLIC
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 48 442 km2
1.3 Population: 5 200 000
1.4 Main climatic and vegetative cones:
Low montane humid forest
Sub- tropical dry forest
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 1 100 000 ha
2.2 Proportion of land under forest: 23%
2.3 Does the country have a written statement of National Forest Policy? NO
2.4 Legislation available to implement policy? NO
2.5 Ownership of forests:
Under State control: 701
Private ownership: 20%
Community ownership: - No effective control: 10%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Fuel wood
Piles for fencing
Poles for power lines
2.7 Staff State Others
2.7.1 Professional: 2 4
2.7.2 Subprofessional (with diploma or
certificate of training): 30 5
2.8 Gross annual budget for forestry: US$ 3 000 000
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area If of plantations at the end of 1978:
2 000 ha
3.1.2 Planned annual target/reforestation:
1 200 ha/year
3.2 Organisation and administration of planting schemes
3.2.1 State Forest Services: 95% Others (specify): 5 mining and private companies
3.3 Principal productor purpose envisaged (for example, taw-wood, posts and stakes,
pulwood, fuelwood, protection, etc.):
3.3.1 Indigenous species:
Mean annual increment (without bark)
Product/ Net Area Rotation at the end of the rotation
Purpose Species (ha) I/ (years) (m3/ha/year)
Protection Pinus
occidental is 1 000 30
II The net area is the gross area of plantations, less the area occupied by roads, paths,
*" firebreaks, buildings and other areas not covered by trees.
- 245 -
3.3.2 Introduced species:
Product/Purpose
Protection and
saw-wood
Fuelwood
Species
Pinus
caribaea
Eucalyptus
robust a
Net area
(ha) I/
800
Rotation
(year)
20
Mean annual increment
(without bark) at the
end of the rotatiqn
(m3/ha/year)
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? NO
4.1.2 Is there a national seed certification system? NO
4.1.3 Are there facilities for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the specied listed
in 3.3.1?
in 3.3.2?
YES
NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? NO
- 246 -
Country Statement: ECUADOR
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 2 745.4 km 2
1.2 Location: longitude 7830 f ; latitude: 00 f
1.3 Population: 7 000 000
1.4 Main climatic and vegetative zones:
From tropical desert to perennial snowfields
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 17 734 000 ha
2.2 Proportion of land under forest: 64.45%
2.3 Does the country have a written statement on National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
Domestic utilisation of forest raw materials and conservation of renewable
natural resources.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
Under State control: 12.41%
Private ownership: 0.12%
Community ownership: 0.02%
No effective control: 87.45%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) :
Saw-wood; logs for triplex and plywood sleepers; posts; stakes; fuelwood and
charcoal .
2.7 Staff State Others
2.7.1 Professional 53 20
2.7.2 Subprofesslonal (with diploma
or certificate of training) 114 33
2.8 Gross annual budget for forestry: US$ 600 000
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1976: 25 425 ha
3.1.2 Planned annual tar get reforest at ion: 5 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 70.0%
Others (specify): 30.0%
3.3 Principal product or purpose envisaged (for example, saw wood, posts and stakes,
pulpwood, fuelwood, protection, etc.):
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 247 -
3.3.1
Indigenous species:
Product/
Species Net area
Rotation
Mean annual
Purpose
(ha I/)
(years)
increment
(without bark)
at the end of
the rotation
(m^ /ha/year)
Saw-wood and
Eucalyptus glo- 17 405
20
18
fuelwood
bulus
Saw-wood
Pinus radiata 5 177
30
14
Saw-wood and
Tectona grandis 646
30
12
plywood
Saw-wood and
Eucalyptus 156
20
16
fuelwood
saligna
Saw-wood and
Eucalyptus 132
20
14
stakes
camaldulensis
Various uses
Various species 129
variable
variable
3.3.2
Introduced species:
Purpose/Product
Species Net area
Rotation
Mean annual
(ha) I/
(years)
increment
(without bark)
at the end of
(m3/ha/year)
Saw-wood and
Cordia alliodora 1 408
25
14
plywood
Conservation
Prosopis juliflora 173
indefinite
10
Saw-wood and
Centrolobium 39
40
14
poles
patinense
Conservation
Pseudosamanea 38
indefinite
18
guachapele
Various
Various species 122
indefinite
variable
4.
SEED AND PLANT SUPPLY
4.1
Service and control
of seed supply
4.1.1
Have seed zones been defined for indigenous species
(i.e. is it common to restrict
the use of seed for
afforestation/reforestation to
the particular
zone in which it
was collected)? NO
- It is collected for local reforestation.
4.1.2
Is there a national
seed certification system? NO
4.1.3
Are there facilities
for storing seed at controlled
temperatures?
NO
4.1.4
Does the supply of seed cover the demand for the species listed in 3.3.1? YES
in 3.3.2? For Eucalyptus globulus and Tectona fcrandis.
5.
TREE IMPROVEMENT
5.1
Does the country have an official tree improvement
programme? NO
I/ The net area is the gross area of plantations, less the area occupied by roads, paths
firebreaks, buildings and other areas not covered by trees.
-248-
Country Statement: GUATEMALA
1. GENERAL GEOGRAPHIC INFORMATION
1.1 Area of country: 108 900 km 2
1.3 Population: 7 045 800
1.4 Main climatic and vegetative zones:
a) South coast: Tropical and sub-tropical forests
b) Western high plateau: Cold, humid and very himid, mixed and coniferous forests.
c) Central region: Conifers in the western cold humid zone, tropical and dry
sub-tropical forests in the east.
d) Eastern high plateau: Cold, dry, conifers and mixed forests.
e) Northern intermediate: Humid and very humid tropical, conifers and hard woods.
f) North (PetSn): Tropical forests of valuable species, very humid.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 3 610 000 ha
2.2 Proportion of land under forest: 39.11
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
a) To conserve the existing forest area;
b) To extend the forest area through afforestation and managed artificial
reforestation; and
c) To encourage the rise of forest industries.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests: Under State control: 100%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Saw-wood (palo bianco, mahogany, cedar, pine);
logs for veneering, for posts and for fuel wood.
2.7 Gross annual budget for forestry: US$ 11 000 000
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978: 3 610 000 ha
3.1.2 Planned annual target reforestation: 12 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 100Z
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.):
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 249 -
3.3.1 Indigenous species:
Product /Purpose
Reforestation
3.3.2 Introduced species:
Product /Purpose
Reforestation
Reforestation
Species
Conifers
Hardwoods
Net area
(ha) I/
Rotation
(years)
Mean annual increment
(without bark) at the
end of the rotation
(m 3 /ha/year)
3 000
9 000
Species
Leucaena
leucocephala
Sesbania
aculeata
Mean annual increment
(without bark) at the
Net area Rotation end of the rotation
(ha) I/ (years) (m3/ha/year)
Recently introduced
Recently introduced
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected? YES - Government Decree of 8 March 1979.
4.1.2 Is there a national seed certification system? YES, BANSEFOR
4.1.3 Are there facilities for storing seed at controlled temperatures? YES, BANSEFOR
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1?
in 3.3.2?
YES
YES
5.
5.1
5.2.2
TREE IMPROVEMENT
Does the country have an official tree improvement programme? NO
Area of seed stands in each of the main species:
Conifers
Hardwoods
1 295 ha
200 ha
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
Conifers (155)
Hardwoods -
5.2.4 Seed orchard (species, area and number of clones or mother trees)*
Leucaena leucocephala
Sesbania aculeata
3 ha (1 000)
1 ha (100)
1J The net area is the gross area of plantations, less the area occupied by roads,
paths, firebreaks, buildings and other areas not covered by trees.
- 250 -
Country Statement: HONDURAS
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 113 000 km 2
1.2 Location: longitude: 85<>; latitude: 14
1.3 Population: 4 million
1.4 Main climatic and vegetative zones: Tropical
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 7.4 million ha
2.2 Proportion of land under forest: 65%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
To preserve for the nation the inestimable flora, fauna and soil resources
in forest areas;
To ensure the protection and improvement of the above;
To rationalize the utilization, industrialization and marketing of forest products.
2.4 Legislation available to implement policy? YES
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Saw-wood, veneers and sleepers, fuelwood.
2.7 Staff State Others
Professional: 76
Subprofessional (with diplc
or certiflcat of training): 196
2.8 Gross annual budget for forestry: US$ 80 million
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.2 Planned annual target area of afforestation/reforestation: 12 600 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 100%
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? NO
4.1.2 Is there a national seed certification system? YES
4.1.3 Are there facilities for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1? YES
in 3.3.2? Very little work
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? YES
- 251 -
5.1.1 If yes, list species Involved:
Pinus oocarpa Llquidambar styraciflua
Pinus carlbaea Dldymopanax morototoni
Pinus pseudostrobus Quercus sp.
Cordla alliodora Leucaena leucocephala
5.1.2 Most important characters to be Improved or bred for:
Shape of stem Basic density
Resin production
Fibre length
5.2 Brief outline of improvement methods already applied:
5.2.1 Species /provenance trials (Indicate in parentheses the number of provenances
being tested) :
Pinus oocarpa (29) Pinus pseudostrobus (18)
Pinus caribaea (26) Cordia alliodora (6)
5.2.2 Area of seed stands in each of the main species:
Not yet decided.
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
Selection has not yet been completed.
5.2.4 Seed orchards (species, area and number of clones or mother trees):
The first one will be established this year; there are no data as yet.
5.2.5 Progeny testing (species and area):
To be started in 1980.
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree improvement in the country, in publications,
reports, etc.
We are trying to publish some experiences.
~ 252 -
Country Statement: NICARAGUA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 130 000 km 2
1.2 Location: longitude: 1045 f - 1505^ latitude: 8330 f - 8720 f
1.3 Population: 2 500 000
1.4 Main climatic and vegetative zones:
Tropical dry forest. Tropical humid forest. Montane and pine wood savanna.
Tropical rain forest, monsoon forest, tropical savanna and sub-tropical montane
climate.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 4 550 000 ha
2.2 Proportion of land under forest: 35Z
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
Nationalization and rational use of forests.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
Under state control: 20%
No effective control: 80%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Saw-wood: 1 000 000 m3/year
2.7 Staff State Others
Professional : 7 3
Subprofessional (with diploma
or certif cate of training) : 12 3
2.8 Gross annual budget for forestry:
US$ 1.5 x 10 6
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978:
1500 ha
3.1.2 Planned annual target /reforestation: 5 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 100%
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.):
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 253 -
3.3.1 Indigenous species:
Product/Purpose
3.32
Species
Timber for pulping Pinus
caribaea
Fuelwood Leucaena sp.
Introduced species:
Product/Purpose
Protection and
fuelwood
Species
Net Area
(ha) I/
3 000
500
Net Area
(ha) I/
Eucalyptus 1 000
Camaldulensis 500
Rotation
(years)
25
4
Rotation
(years)
Mean annual increment
(without bark) at the
end of the rotation
(m 3 /ha/year)
4
30
Mean annual Increment
(without bark) at the
end of the rotation
(m 3 /ha/year)
30
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is It common to restrict
the use of seed for afforestation/reforestation to the particular zone in which
it was collected)? NO
4.1.2 Is there a national seed certification system? NO
4.1.3 Are there facilties for storing seed at controlled temperatures? NO
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1?
in 3.3.2?
YES
NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? NO
I/ The net area is the gross area of plantations, less the area occupied by roads,
~ paths, firebreaks, buildings and other areas not covered by trees.
- 254 -
Country Statement: PANAMA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 77 082 km 2
1.2 Location: Longitude: 779' - 833'; latitude: 712' - 938' N
1.3 Population 1 881 400
1.4 Main climatic and vegetative zones:
Dry tropical forest, humid tropical forest, very humid tropical forest, very humid
premontane forest, premontane rainforest, very humid low montane forest.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 3 225 900 ha
2.2 Proportion of land under forest: 50%
2.3 Doest the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
To reduce importation by developing local industries, promoting local products,
with the intention of remedying shortages in local supplies of the raw material
necessary for production of wood and wood products.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests: Under state control: 100%
2.6 Principal forest products (for example, saw-wool, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Furniture, posts, beams, saw-wood, logs for sleepers, plywood, lumber.
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.2 Planned annual target area of afforestation/reforestation: 10 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 100%
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuel wood, protection, etc.):
3*3.1 Indigenous species:
Mean annual increment (without bark)
Product/
Purpose
Furniture
and beams
Net Area Rotation
Species (ha) I/ (years)
at the end of the rotation
(m 3 /ha/year)
Cedrela
odorata
Forest-
plots
Bombacopsis
quinata
Fence
posts
Cordia
alliodora
Wood
Tabebuia
pentaphylla
Wood
Tabebuia
guayacan
I/ The net area in the gross area of plantations, less the area occupied by roads, paths
firebreaks, buildings and other areas not covered by trees*
- 255 -
3.3.2 Introduced species:
4.
4.1
4.1.1
4.1.2
4.1.3
4.1.4
5.
5.1
Product/
Purpose
Species
Mean annual increment (without bark)
Net Area Rotation at the end of the rotation
(ha) I/ (years) (m3/ha/year)
Posts, wood Tectona
grandis
Wood
Gmelina
arborea
Reforestation Pinus
caribaea
5.41
7.7
2 400
Wood
Wood
Anthocephalus
cadamba 1 . 5
Hibiscus
elatus
0,95
SEED AND PLANT SUPPLY
Service and control of seed supply
Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? YES
Is there a national seed certification system? YES
Are there facilities for storing seed at controlled temperatures? YES
Does the supply of seed cover the demand for the species listed in 3.3.1? 65%
in 3.3.2? 50%
TREE IMPROVEMENT
Does the country have an official tree improvement programme? NO
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
"" firebreaks, buildings and other areas not covered by trees.
- 256 -
Country Statement: PARAGUAY
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 406 752 km 2
1.2 Location: Longitude: 54 - 63W; latitude: 19 - 28S
1.3 Population: 3 500 000
1.4 Main climatic and vegetative zones:
Sub-tropical temperate clirnate
Vegetation: thorny steppe and dry forest (western region) - HOLDRIDGE
Humid forest (eastern region) - HOLDRIDGE
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 6 000 000 ha
2.2 Proportion of land under forest: 45%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
a) Management plan; b) Reforestation plan; c) National parks plan; d) Five-year
plan for industry.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
Under State control: 10%
Private ownership: 90%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey,
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood):
Sawn wood; veneer logs, logs for sleepers, posts, poles and stakes, fuelwood and
charcoal.
2.7 Staff State Others
Professional: 32 10
Subprofessional (with diploma
or certificate of training): 55 10
2.8 Gross annual budget for Forestry:
US$ 430 000
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantation at the end of 1978: 5 000 ha
3.1.2 Planned annual target area of afforestation/reforestation: 1 500 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: - Others (specify): 100% (private)
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 257 -
3*3.2 Introduced species: w . .
Mean annual increment
(without bark) at the
Product/ Net Area Rotation end of the rotation
Purpose Species (ha) J7 (years) (m3 /ha/year)
Pulpwood Pinus elliottii 4 500 25 21.1
Pulpvood Pinus taeda 30 25 21.3
Saw- wood Araucaria angustifolia 3 30 15.8
Pulpwood Cupressus japonica 2 25 9.2
Note: Results in stands of 13 years, with 20 percent thinning at 9 years.
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it conmon to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? NO
4.1.2 Is there a national seed certification system? NO
4.1.3 Are there facilities for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the species listed in 3.3.1? NO
in 3.3.2? NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme? NO
5.1.2 Most important characters to be improved or bred for:
a) straight bole
b) rapid growth
c) physical-mechanical aptitudes
d) resistance to pests, diseases and adverse climatic conditions
5.2 Brief outline of improvement methods already applied:
5.2.1 Species /provenance trials (indicate in parentheses the number of provenances
being tested). Various exotic species of unknown specific origin have been
introduced .
Pinus elliottii (2) P. caribaea var. caribaea (2)
P. taeda (2) P. caribaea var. hondurensis (1)
Araucaria angustifolia (1) P. palustris (1)
Cupressus japonica (1) Eucalyptus spp. (12 spp.)
5.2.6 Other methods of improvement (specify):
Note: In the second half of 1979, needs of various exotic species (conifers, and
hardwoods) were introduced; collection of seeds of native species was also started,
all of them tested but without final results.
7. REFERENCES TO TREE IMPROVEMENT
Note: We have no official reports on forest improvement, except for some publica-
tions of partial results relating to the introduction of species, adaptability, etc.
I/ H ntt area it the gross area of plantations, less the area occupied by roads, paths,
~ firebreaks, buildings and other areas not covered by trees.
- 258 -
Country Statement: PERU
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 1 285 215 km 2
1.2 Location: Longitude: 70 - 815 latitude: - 18S
1.3 Population: 20 000 000
1*4 Main climatic and vegetative zones:
Coastal region: dd-PT, ds-PT, dp-PT (ppx: 0-200 mm; T.x: 21 C)
Mountain region: bb-MT, ee-MBT, bmh-MT (ppx: 250-1 400 mm; T.x: 14)
Forest region: bh-T, bmh-T (ppx: 900-4 000 mm; T.x: 24 C)
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 74 120 000 ha
2.2 Proportion of land under forest: 60%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
To meet the national demand for wood and forest products.
To earmark land suitable for forests through reforestation to promote social and
economic development.
To lay the basis for optimum utilization of forests and wildlife resources.
2.5 Ownership of forests:
Under State control: 99.87.
Private ownership: 0.047.
Community ownership: 0.16%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer
logs, logs for sleepers, posts, piles and stakes, pulpwood):
Saw-wood, veneer logs, plywood, parquet, sleepers, posts, timbers for mines, fuel-
wood, charcoal.
2.7 Staff State Others
Professional: 240
Subprofessional (with diploma or
certificate of training): 130
2.8 Gross annual budget for forestry:
US$ 6 000 000 approximately
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area y of plantations at the end of 1978: 120 000 ha
3.1.2 Planned annual target/reforestation: 15 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State forest Services: 98%
Others (specify): Private ownership: 2%
I/ The net area in the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 259 -
3.3
Principal
product or purpose envisaged 'for example, saw-wood, posts and stakes,
pulpwood,
fuelwood, protection, etc.)
3.3.1
Indigenous
species:
Product/
Net Area
Purpose
Species (ha) I/
Saw-wood
Cedrela sp. )
and
Swietenia ) 1 800
Plywood
Cedrelinga )
Charcoal,
Prosopis sp. )
fuelwood,
Tecoma sp. ) 4 000
fruit,
Lexopterygium )
parquet
3.3.2
Introduced
Mean annual increment (without bark)
Product/
Net Area Rotation at the end of the rotation
Purpose
Species (ha) I/ (ha) I/ (m3/ha/year)
Saw-wood
Eucalyptus
posts,
globulus
sleepers.
108 000 20 18-20
fuelwood,
protection
Sawn wood
Pinus radiata 3 000 20 14-18
4.
4.1
4.1.1
4.1.2
4.1.3
4.1.4
5.
5.1
5.1.1
5.1.2
5.2 ,
5.2.2
Sawn wood
Cupressus sp )
Pinus caribaea )
Eucalyptus )
grandis )
E. saligna )
200
20
12-14
SEED AND PLANT SUPPLY
Service and control of seed supply
Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? YES
Is there a national seed certification system? YES
Are there facilities for storing seed at controlled temperatures? YES
Does the supply of seed cover the demand for the species listed in 3.3.1? YES
in 3.3.2? NO
TREE IMPROVEMENT
Does the country have an official tree improvement progr
? YES
If yes, list species involved:
Pinus radiata
? greggii
P. patula
P. elliottii
Eucalyptus globulus
E. grandis
Most important characters to be improved or bred for:
Rapid growth
Straight trunk
Resistance
Brief outline of improvement methods already applied:
Area of seed stand in each of the main species: Pinus radiata (20 ha)
I/ The net area is the gross area of plantations, less the area occupied by roads,
~ paths, firebreaks, buildings and other areas not covered by trees.
- 260 -
Country Statement: URUGUAY
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 187 000 km*
1.2 Location: longitude: 54-58W; latitude: 30-35S
1.3 Population: 3 000 000
1.4 Main climatic and vegative zones:
Due to its small size, the country does not have marked climatic or vegetative
variations; the climate is mild temperate and the vegetation mainly grasslands.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 746 000 ha
2.2 Proportion of land under forest: 3.97.
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
Meeting the internal demand for forest products (sawn wood, pulp and paper,
fuel-wood, lumber, etc.)
Export of production surpluses
Concentration of forest stands
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
Under State control: 3%
Private ownership: 97%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer
logs, logs for sleepers, posts, piles and stakes, pulpwood):
Fuelwood : 1 200 000 n>3
Saw and board wood : 150 000 m^
Pulpwood : 115 000 to?
Veneer wood : 12 000 m 3
2.7 Staff State Others
Professional: 32 60
Subprofessional (with diploma or
certificate of training): 8 20
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area I/ of plantations at the end of 1978: 157 500 ha
3.1.2 Planned annual target area of afforestion/reforestation: 5 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 25%
Others (specif iy): Private ownership: 75%
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuel wood, protection, etc.):
I/ The net area is the gross area of plantations, leas the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by tree\
- 261 -
3.3.2 Introduced species:
4.
4.1
4.1.1
4.1.2
4.1.3
4.1.4
5.
5.1
5.1.1
5.1.2
5.2
5.2.1
Mean annual increment (without bark)
Product/
Net area
Rotation
at the end of the rotation ,
Purpose
Species (ha) \J
(years)
(m^/ha/year)
Saw-wood,
P. elliottii,
pulp and
P. taeda, P. 25 000
20
18
board
pinaster
Saw-wood,
Eucalyptus
Fuelwood,
umbel lata, E.
pulp
camaldulensis, 110 000
20
22
board
E. globulus,
E. grandis
Crates and
Salicaceas
boxes, saw-
8 000
18
16
wood
SEED AND PLANT SUPPLY
Service and control of seed supply
Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? NO
Is there a national seed certification system?
Being established.
Are there facilities for storing seed at controlled temperatures? NO
Does the supply of seed cover the demand for the species listed
in 3.3.1?
in 3.3.2? NO
TREE IMPROVEMENT
Does the country have an official tree improvement programme? YES
If yes, list species involved:
Pinus elliottee, P. taeda
P. patula, P. pinaster,
P. radiata
Eucalyptus grandis
E. globulus, E. smithii
E. maidenii, E. resinifera,
E. Bosistoana, E. umbel lata
Salix sp., Populus sp.
Platanus occidentalis
Most important characters to be improved or bred for:
Growth Resistance to climatic factors
Conformation and disease and pests
Timber quality
Brief outline of improvement methods already applied:
Species/provenance trials (indicate in parentheses the number of provenances
being tested):
Pinus taeda (11) Eucalyptus umbellata (15)
Pinus elliottii (5) Eucalyptus maidenii (3)
Pinus patula (3)
Eucalyptus grandis (15)
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
~ firebreaks, buildings and other areas not covered by trees.
- 262 -
5.2.2 Area of seed stands in each of the main species:
Pinus taeda (20 ha) Eucalyptus grandis (1,5 ha)
Pinus elliottii (15 ha)
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Pinus taeda: 4 ha, 12 mother trees
Pinus elliottii: 4 ha, 19 mother trees
5.2.5 Progeny testing (species and area):
Eucalyptus grandis: 0.2 ha
5.2.6 Other methods of improvement (specify):
Clonal selections are being made on Salix sp., Populus sp. and Platanus occidentalis
6. SUCCESSFUL METHODS IN VEGETATIVE PROPAGATION FOR MAIN SPECIES:
Species Method % of success
Salis sp., Populus sp. Untreated cutting 90-100%
Platanus occidentalis Untreated cutting 85-95%
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree improvement in the country, in publications
reports, etc.
Krall, J. adaptabilidad de confferas de Norteamerica plantadas en el Uruguay y su
susceptibilidad a insectos y enfermedades. Bole tin del Departamento Forestal N 16.
Krall, J. Populus: datos sobre fenologfa y crecimiento en el Uruguay. Boletfn del
departament Forestal NO 17.
Krall, J. Fundamentos para nuevas introduce iones de Eucalyptus en el Uruguay.
Facultad de Agronomia, Boletfn N 17.
- 263 -
Country Statement: VENEZUELA
1. GENERAL GEOGRAPHICAL INFORMATION
1.1 Area of country: 913 050 km 2
1.2 Location: Longitude: 5948 f - 73ll t 49"W; latitude: 043' - 1211'46 M N
1.3 Population: 15 000 000
1.4 Main climatic and vegetative zones:
Dry tropical forest, humid tropical forest, very dry tropical forest, very humid
premontane forest.
2. FORESTS AND NATIONAL FOREST POLICY
2.1 Area of forests: 47 971 000 ha
2.2 Proportion of land under forest: 53%
2.3 Does the country have a written statement of National Forest Policy? YES
2.3.1 If a National Policy exists, what are the main objectives stated in it?
The objectives follow the major guidelines for the three classes of forest:
protective, productive and the secondary functions such an recreation, research,
education, etc.
2.4 Legislation available to implement policy? YES
2.5 Ownership of forests:
Under State control: 82%
Private ownership: 18%
2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer
logs, logs for sleepers, posts, piles and stakes, pulpwood):
Saw-wood, veneer logs, particle board, sleepers, posts.
2.7 Staff State Others
Professional: 160 294
Subprofessional (with diploma or
certificate of training):
2.8 Gross annual budget for forestry:
US$ 20 000 000
3. AFFORESTATION AND REFORESTATION
3.1 Areas
3.1.1 Net total area \j of plantations at the end of 1978: 101 850 ha
3.1.2 Planned annual target area of afforestation/ reforestation: 32 000 ha/year
3.2 Organization and administration of planting schemes
3.2.1 State Forest Services: 85% Others (specify): Combined State and Private
ownership: 107.
Private ownership: 5%
3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes,
pulpwood, fuelwood, protection, etc.):
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 264 -
3.3.1 Indigenous species:
Product/
Purpose
Species
Mean annual increment (with-
Net Area Rotation out bark) at the end of the
(ha) I/ (years) rotation (m 3 /ha/year)
Mean annual increment (with-
out bark) at the end of the
rotation (mVha/year)
20
10
Saw-wood Anacardium excel sa,)
Swietenia macro- )
phylla, Oedrela )
odorata, Cordia ) 2 800
alliodora, Cordia )
apurensis, Tabebuia)
rosea )
3.3.2 Introduced species:
Product/ Net Area Rotation
Purpose Species (ha) I/ (years)
Pulp Eucalypts 3 000 6
Pulp Caribbean Pine 101 850 12
Saw-wood Teak 620 30
Protection Pine, Eucalypts
and others
4. SEED AND PLANT SUPPLY
4.1 Service and control of seed supply
4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict
the use of seed for afforestation/reforestation to the particular zone in which it
was collected)? YES
4.1.2 Is there a national seed certification system? YES
4.1.3 Are there facilities for storing seed at controlled temperatures? YES
4.1.4 Does the supply of seed cover the demand for the species listed
in 3.3.1? YES in 3.3.2? NO
5. TREE IMPROVEMENT
5.1 Does the country have an official tree improvement programme?
In-depth tree breeding studies are done only in autonomous research organizations,
or institutes: there is no State programme at national level.
5.1.1 If yes, list species involved:
Pinus caribaea var. hondurensis
Bombacopsis quinata
5.1.2 Most important characters to be improved or bred for:
Straightness of stem, bifurcations, anomalies, height, wood density and volume
(Caribbean pine). Buttresses, bifurcation in Bombacopsis quinata.
5.2 Brief outline of improvement methods already applied:
5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being
tested):
Genero Pinus (32 spp.)
Indigenous hardwoods (10 spp.)
Exotic hardwoods (25 spp.)
I/ The net area is the gross area of plantations, less the area occupied by roads, paths,
firebreaks, buildings and other areas not covered by trees.
- 265 -
5.2.2 Area of seed stands in each of the main species:
Pinus caribaea var. hondurensis (520 ha) - in process of establishment (CVG-CONARE)
5.2.3 Plus trees of main species (indicate in parentheses the number of trees):
There are 40 Pinus caribaea trees selected as candidates
5.2.4 Seed orchards (species, area and number of clones or mother trees):
Bombacopsis quinata
Tabebuia rosea
Pinus caribaea
5.2.5 Progeny testing (species and area):
Bombacopsis quinata
Tabebuia rosea
5.2.6 Other methods of improvement (specify):
Controlled and open pollination of Bombacopsis quinata
6. SUCCESSFUL METHODS IN VEGETATIVE PROPAGATION FOR MAIN SPECIES:
Species Method % of success
Bombacopsis quinata Grafting and cutting 90
Cedrela odorata Grafting
Tubebuia rosea Grafting
Pinus caribaea Grafting, aerial shoots 30-50
7. REFERENCES TO TREE IMPROVEMENT
Please list the references to tree improvement in the country, in publications,
reports, etc.
Quijada and Gutierrez. Estudio sobre la propagacion vegetativa de especies
forestales venezolanas.
Mendez and Luis. Etapas preliminares para el establecimiento de un rodal semi Hero
en Chaguaras Estado Monagas.
Norman Smith. Seleccion de arboles en Cachipo para establecer un huerto semillero
de Pinus caribaea. (Tree selection in Cachipo for the establishment of a Pinus
caribaea seed orchard).
Appendix VII
BIBLIOGRAPHY
1. Publications distributed to all participants
a. Statistics, experimental design, selection of species/provenances
Freese, F., 1978. Me*todos estadfsticos eleroentales para tecnicos forestales.
Servicio Forestal, Departamento de Agricultura de los EE.UU. de America.
Manual de Agricultura Num. 317. Centre Regional de Ayuda Tecnica, AID.
Mexico/BsAs.
Burley, J. & Wood, P.J., 1979. (Editors). Manual sobre investigaciones de especies
y procedencias con referencia especial a los tr6picos. CFI Tropical Forestry
Papers no. 10 & 10A (Edici6n Espafiola). Oxford, Reino Unido.
b. Conservation of forest genetic resources
Roche, L. , The met ho do logy of conservation of forest genetic resources* Report on
a Pilot Study. FO:MISC/75/8. FAO/UNEP, Rome.
FAO, 1977. Report of the Fourth Session of the FAO Panel of Experts on Forest
Genetic Resources. FO:FGR/4/Rep. FAO, Rome.
FAO, 1975. Forest genetic resources information No. 4 (Global programme for improved
use of forest genetic resources). Forestry Occasional Paper 1975/1. FAO, Rome.
FAO, 1978. Forest genetic resources information No. 8. Forestry Occasional Paper
1978/2. FAO, Rome.
c. Reports on conferences and technical consultations
FAO, 1978. Genetics, Selected papers from the Third World Consultation on
Improvement of Forest Trees. Unasylva 30 (119/120). FAO, Rome.
d. Forest tree improvement
Faulkner, R., 1975. Seed Orchards. Forestry Commission Bulletin No. 54. London,
U.K. 1975.
Koenig, A. & Melchior, G.H., 1978. Propagacion vegetativa en arboles forestales.
FAO/PNUD/ COL/74/005, P.I.F. No. 9. INDERENA, Bogota.
e. Collection of forest seeds for storage and treatment
Greaves, A. y 1978. Descriptions of seed sources and collection for provenances of
Pinus caribaea. Commonwealth Forestry Institute, Oxford. Tropical Forestry
Papers No. 12. Oxford, U.K.
Bonner, F.T., 1977. Equipment and Supplies for Collecting, Processing, Storing and
Testing Forest Tree Seed. USDA Forest Service, General Technical Report SO-13.
Southern Forest Expt. Station, New Orleans, Louisiana.
Yeatroan, C.W. & Nieman, T.C., 1978. Safe tree climbing in forest management.
Petawawa Forest Experiment Station, Chalk River, Ontario. Forestry Technical
Report 24.
FAO 1977. Manual I, sobre elcccion de rodales para la recolecci6n de seraillas
forestales. INAFOR-BANSEFOR-FAO. FAO/TCP GUA-6/01-T, Guatemala.
FAO, 1978. Manual II, para la recolecci6n de semi 11 as forestales. INAFOR-BANSEFOR-
FAO/TCP GUA-6/01-T, Guatemala.
- 267 -
van Dijk, K. , Venegas Tovar, L. , & Melchior, G.H. , 1978. El suministro de serai 11 as
como base de reforestaciones en Colombia. FAO/PNUD COL/74/005, P.l.F. No. 13.
INDERENA, Bogota.
Gordon, A.W. , 1979. Uso y abastecimiento de semillas forestales en Chile. FAO/PNUD,
DP/CHI. 76/003, Documento de Trabajo No. 16. Santiago de Chile.
f . Information on specific species
g. Reforestation, forest plantations
FAO, 1978. Establishment Techniques for Forest Plantations. FAO Forestry Paper No. 8.
FAO, Rome.
Navarro, G. & Molina, R.J., & Montero de B., J.L., 1975. Tecnicas de forestaci6n.
ICONA, Monograffas No. 9. (2a edition). Madrid, Spain.
Sommer, A. & Dow, T., 1978. Compilation of indicative growth and yield data on fast-
growing exotic tree species planted in tropical and sub-tropical regions. FO: MISC/
78/11. FAO, Rome.
Lanley, J.P. & Clement, J. Present and Future Forest and Plantation Areas in the
Tropics. FO/MISC/79/1. FAO, Rome.
h. General
FAO, 1978. Publications on sale. FAO, Rome.
2. Publications available for reference
a. Statistics, experimental design, selection of species/provenances
FAO, 1970. Majoramiento de arboles y biometra, Cuba. FAO: SF/CUB 3, Informe
Teen i co No. 1. FAO, Rome.
FAO, 1972. Genetica y biometra, Cuba. FAO:SF/CUB 3, Informe TScnico No. 3. FAO,
Rome.
b. Conservation of forest genetic resources
FAO, 1973-78. Forest genetic resources information, Nos. 1-8. Forestry Occasional
Paper. FAO, Rome.
Lucas, G. & Synge, H. , 1978. (Editores). The IUCN Plant Red Data Book. IUCN/WWF,
Morges, Switzerland.
Frankel, O.K. & Bennett, E., 1970. (Editores). Genetic resources in plants: their
exploration and conservation. IBP Handbook No. 11. Blackwell Scientific
Publications, London, U.K.
c. Reports on conferences and technical consultations
Anon, 1978. Actas de la tercera consulta mundial FAO/IUFRO sobre la mejora de
arboles forestales, I-III. CSIRO, Canberra, Australia.
Burley, L. & Nikles, D.G., 1972-73. (Editores). Selection and breeding to improve
some tropical conifers. Based on papers submitted to a Symposium of IUFRO Working
Groups on Breeding of Tropical and sub-tropical Species and on Quantitative
Genetics applied to forest trees, held in Gainesville, Florida, 1971. CFI,
Oxford, U.K.
Burley, J. & Nikles, D.G. Tropical provenance and progeny research and international
cooperation. Proceedings of a joint Meeting of IUFRO Working Parties S.2.02.8 and
82. 03.1, held in Nairobi, Kenya 1973. CFI. Oxford, U.K. 1973.
- 268 -
Nikles, D.G., Burley, J. & Barnes, R.D., 1978. Progress and problems of genetic
improvement of tropical forest trees. Proceeding of a joint workshop of IUFRO
Working Parties S2.02.8 and S2.03.1, held in Brisbane, Australia, 1977. CFI,
Oxford, U.K.
FAO, 1963. Proceedings of the World Consultation on Forest Genetics and Tree
Improvement, Stockholm. FAO, Rome.
FAO, 1969. Proceedings of the Second World Consultation on Forest Tree Improvement,
Washington, FAO, Rome.
FAO, 1964. FAO/IUFRO meeting on forest genetics. Unasylva 18 (2-3), Nos. 73-74.
FAO, Rome.
FAO, 1970. FAO/IUFRO Consultation on frest tree breeding. Unasylva 24 (2-3)
Nos. 97-98. FAO, Rome.
d. Forest tree improvement
Wright, J.W. Genetics of Forest Tree Improvement. FAO Forestry and Forest Products
Studies No. 16. FAO, Rome.
Allard, L.W. (trad. J.L. Montoya). Principios de la mejora genetica de las plantas.
Editorial Omega S.A., Barcelona.
Baldwin, R.E. Genetica elemental. Editorial Limusa, Mexico D.F., 1976.
Hartmann, H.T. & Kester, D.E., 197- . Propagacion de plantas: teorfa y practice.
Editorial Continental, Mexico.
Stern, K. & Roche, L., 1974. Genetics of Forest Ecosystems. Ecological Studies
No. 6. Chapman & Hall Ltd., London, U.K.
Dorman, K.W. , 1976. The Genetics and breeding of southern pines. USDA Agric.
Handbook No. 471 USDA, Forest Service. Washington D.C.
FAO, 1974. Report on the FAO /DAN I DA Training Course on Forest Tree Improvement,
Kenya 1973. FAO, Rome.
ADAA, 1978. Selected reference papers. International Training Course in Forest Tree
Breeding, Canberra 1977. Australian Development Assistance Agency, Australia.
Burdon, R., 1978. Mejoramiento genetico forestal en Chile. FAO/PNUD DP/CHI/76/003,
Documento de Trabajo No. 11. Santiago de Chile.
Melchior, G.H., 1977. Bases y posibilidades del mejoramiento genetico de arboles
forestales en Colombia. FAO/PNUD COL/74/005, P.I.F. No. 4. INDERENA, Bogota.
Melchior, G.H., 1977. Programa preliminar de un ensayo de procedencia de Cordia
alliodora, Cupressus lusitanica y otras especies nativas y ex6ticas. FAO/PNUD
COL/74/005, P.I.F. No. 7. INDERENA, Bogota".
Shelbourne, C.J.A., 1969. Tree Breeding Methods. New Zealand Forest Service.
Technical Development Paper No. 55. Wellington, N.Z.
e. Collection of forest seeds for storage and treatment
FAO, 1975* Report on the FAO /DAN I A Training Course on Forest Seed Collection and
Handling, I-II, Thailand 1975. FAO, Rome.
1STA, 197- . Reglas Internacionales para ensayos de semi lias. Librerfa Agrfcola,
Madrid, Espafia.
Catalan Bachiller, G. t 1977. Semi 11 as de Arboles y arbustos forestales. Min. Agric.
(1CONA) Madrid.
- 269 -
Anon, 1974* Seeds of Woody Plants in the United States. USDA Agric. Handbook No. 450
USDA, Forest Service. Washington D.C.
FAO, 1975. Forest tree seed directory. FAO, Rome.
1BPGR, 1976. Report of IBPGR Working Group on engineering, design and cost aspects
of long-term seed storage facilities. International Board for Plant Genetic
Resources. FAO, Rome.
Lowman, B.J., 1975. Equipment for processing small seedlots. Catalog No. 7524-2505.
USDA Forest Service, Missoula, Montana.
FAO, 1968. Forest tree seed notes. I. Arid areas. II. Humid tropics. FAO Forestry
Development Paper No. 5. FAO, Rome.
Purtich, G.S., 1972.* Cone production in conifers. Canadian Forestry Service,
BC - X - 65. Department of the Environment. Victoria, British Columbia.
Sarvas, R. , 1962. Investigations on the flowering and seed crop of Pinus
silvestris L. Com. Inst. For. Fenniae 53.4. Helsinki.
Melchior, G.H. & Venegas T.L., 1978. Propuesta para asegurar el suministro de
semillas de Eucalyptus globulus en calidad comercial y geneticamente mejoradas.
FAO/PNUD COL/74/005, P.I.F. No. 14. INDERENA, Bogota.
FAO, 1978. Centro de Semillas forestales, Paraguay. Estudio de Factabilidad
Tecnica para su establecimiento, organizacion y funcionamiento. FAO/SFN/TCP
6/PAR/02-T. Asuncion, Paraguay.
Bramlett, D.L., 1978. e_t aj^. Cone analysis of Southern Pines: a Guidebook. USDA,
Forest Service General Technical Report No. SE-13. Washington D.C.
f . Information on specific species
FAO, 1973/74/75. Annotated bibliographies on Pinus patula, Cupressus lusitanica,
and Pinus elliottii. FAO, Rome.
Carruyo, L.J., 1976. Bibliograf fas sobre Pinus caribaea (1938-72; 1973-75), I - II.
Laboratorio Nacional de Productos Forestales, Publicaciones Tecnicas 76-01-02,
01-03 76. Merida, Venezuela.
CFI, Oxford, "Fast Growing Timber Trees of the Lowland Tropics 11 , numbers:
1. Gmelina arborea (1968)
2. Cedrela odorata (1968)
3.' The araucarias (1968)
4. Pinus merkusii (1968)
5. Terminalia ivorensis (1971)
6. Pinus caribaea (1973)
CFI, Oxford, "Tropical Forestry Papers", numbers:
7. Pinus patula (1975)
8. Eucalyptus camaldulensis (bibliography) (1975)
9. Pinus kesiya (1979)
10. Agathis (1977)
US Department of Agriculture, Forest Service; Research Papers:
WO - 19 (1973): Genetics of loblolly pine
WO - 20 (1974): Genetics of slash pine
WO - 35 (1978): Genetics of douglas fir
Gibson, I.A.S., 1975. Diseases of forest trees widely planted as exotics in the
tropics and Southern hemisphere. I. Commonwealth Mycological Institute/
Commonwealth Forestry Institute, Oxford, U.K.
- 270 -
Gibson, I.A.S., 1979. Diseases of forest trees widely planted in the tropics and
Southern hemisphere. II. The genus Pinus. CM1/CFI, Oxford, U.K.
Greaves, A., 1979. Descriptions of seed sources and collections for provenances of
Pinus oocarpa. CF1 Tropical Forestry Papers No. 13. Oxford, U.K.
Rowland, P., 1979. Pericopsis elata (Af romorsia) . A summary of silvicultural
knowledge. CFI, Oxford, U.K.
Anon, 1976. Leucaena - promising forage and tree crop for the tropics. National
Academy of Sciences, Washington D.C.
Anon, 1976. Under-exploited tropical plants with promising economic value. National
Academy of Sciences, Washington D.C.
Venegas T.L., 1978. Distribucion de once especies forestales en Colombia. FAO/PNUD
COL/ 74/005, P.I.F. No. 11. INDERENA, Bogota.
Melchior, G.H., 1977. Propuesta para mejorar la produccion del Inchi por metodos
silviculturales y geneticos. FAO/PNUD COL/74/005. P.I.F. No. 8. INDERENA, Bogota.
FAO, 1971. Insectos que atacan a Pinus caribaea en el noreste de Nicaragua.
FAO/SF/NIC 9, Informe Tecnico No.l. FAO, Rome.
Ntima, O.O., 1971. Los Araucarias. Boll, 36-37, IFLAIC, Merida, Venezuela,
FAO, 1980. Populus spp. Handbook on Plantation and Utilization. FAO, Rome (in press).
FAO, 1980. Eucalyptus spp. Handbook on Plantation and Utilization. FAO, Rome
(in press).
g. Reforestation, forest plantations
FAO, 1978. Choice of tree species. FAO Forestry Development Paper No. 13 (2nd Ed.).
FAO, Rome.
FAO, 1974. A manual on the planning of man-made forests. Fo:MISC/73/22. FAO, Rome.
Paul, D.K., 1972. A handbook of nursery practice for Pinus caribaea var. hondurensis
and other conifers in West Malaysia. FO:SF/MAL 12, Working Paper No. 19,
Kuala Lumpur.
FAO, 1960. Forest Plantation Practices in Latin America. Department of Forestry, FAO.
FAO. Dev. Paper No. 15. FAO, Rome.
Cozzo, D., 1976. Tecnologfa de la forestaci6n en Argentina y America Latina. Ed.
Hemisferio Sur, Buenos Aires.
FAO, 1975. Tree planting practices in African savannas. FAO Forestry Development
Paper No. 19* FAO, Rome.
FAO, 1977. Savanna Afforestation in Africa. FAO Forestry Paper No. 11. FAO, Rome
FAO, 1978. Forestry for local community development. FAO Forestry Paper No. 7.
FAO, Rome.
FAO, 1979. Forestry for rural communities. FAO, Rome.
FAO. 1967. FAO World Symposium on Manroade Forests and their Industrial Importance.
Unasylva 21 (3-4), Nos. 86-87. FAO, Rome.
Anon, 1978. Tropical Forest ecosystems: a state-of-knowledge report prepared by
Unesco/UNEP/FAO. Natural Resources Research XIV, Unesco, Paris.
- 271 -
APPENDIX VIII
CERTIFICATE OF PARTICIPATION PRESENTED TO PARTICIPANTS
CURSO de CAPACITACION
FAO/DANIDA
sobrela
MEJORA GENETICA de ARBOLES FORESTALES
Con la presente se certified que
participo en el Curso arriba mencionado
celebrado en
MMda, Venezuela
del 14 de enero al 2 defebrero de 1980
MMda, Venezuela
2 de febrero de 1980
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Open Library
American Libraries
TV News
Understanding 9/11